Something is hilariously off here: Why should I pay $10 and be forced to use it by the end of the month, while I can pay $10 and have it last as long as I want?
is $10 Pro monthly subscription a pre-requisite before i can purchase $10 in API credits?
PS: i would have loved if I can directly buy $10 in credits and be free to spend it as quickly or as leisurly as I want -- without any monthly expiry or fixed recurring payments
How would that be? They are already charging as much as the underlying providers. They can hardly expect to have any customers if they are charging more.
Enterprise sales will be the answer. Microsoft will have some story that convinces an exec eight levels up the org chart from the normal users that this is an essential product they need to overpay for. Given their existing relationshipsand immense sales team they'll probably have success.
That story is data governance. Corporate already have a data-agreement with MS, storing all their data there. Github copilot is covered by that, while a individual agreement with e.g. anthropic needs lawers involved.
It’s precisely this, and, to be fair, it’s a rational approach given a Data Security Exhibit starts at 6 weeks and can hit 6 months to complete. That being said, I work with regulated data, so YMMV.
Plus if you're in government you have procurement to deal with. You already have an Enterprise deal with Microsoft so you don't have to go through any of that rigmarole.
Microsoft is simply the default answer for most large corporations. Getting access to some Microsoft subscription is very easy, because of the existing framework agreements, Microsoft providing any and all compliance slopuments needed and already being pre-cleared for corporate data etc. Meanwhile trying to use another provider (e.g. Anthropic) would be a one year endeavor, minimum.
They list the price 900% higher and give a 90% discount to enterprises who also use teams, outlook, office or even windows if they're desperate. Then that becomes a deal so good that enterprises can't afford not to take it!
e.g. if on an annual plan? 0x will be gone, but there are okay 1x and 0.3x models left. I am pretty much curious how the early may test invoicing will look like. current setup of tools etc. is way too chatty eats up 1+M token per PRU easily. not sure how much is cached.
Only reason to keep it is if you like their UX and auto-complete. Everything else is on pay per use and if you don't use all of it (good luck with the 5 hour and week caps) you have just paid more for the auto-complete
The deal is really pretty much garbage now and I believe that is the intent.
Can you elaborate please? I still clone my private repos and work on them using OpenRouter Credits and OpenCode (or Copilot itself: It supports OpenRouter BYOK). No?
I have to wonder if it's because of how many Enterprise customers they have who have standardized on Github Copilot and gotten it through the gauntlet of legal approvals etc.
This rings true to me as someone who's worked at a few large corps like this. A price hike does not change things when there is a mandate to use MS products over other vendors.
We're putting other providers through the gauntlet. An M4 Studio or two running the latest Qwen3 or whatever counts for state of the art in open models is also looking a little more viable all the time.
They can be super complementary! Open weight models can be your everyday standard goto, and frontier models for the harder and bigger tasks.
Having some open weight deployment or vendor is also a good thing, because you may have domain specific tasks where you can get better results on domain specific problems with a quick finetune.
Unsloth makes it particularly easy. Open weight LLMs are incredibly powerful building blocks.
Bingo. Ghcp is the only allowed llm solution at our big well known semiconductor corp. It'll take years to get approval for anything else. We're stuck with it and will pay whatever price we have to.
I am a bit confused by the separation between VSCode and Copilot. If I cancel my Pro+ subscription, can I still use Copilot with my own OpenRouter key?
I'm wondering if they're basically saying they're going to give $10/month free API credits to students and open source maintainers and so on... while otherwise getting out of the consumer portion of this space.
they're downsizing free github copilot pro for open source maintainers. At the very least, it looks like small open source projects got their free copilot pro cut off
I was kinda of disappointed when that happened earlier this month, but not as much now after seeing this change. My primary use had been trying out some of the newer Anthropic and OpenAI models, which probably would have burned through $10 worth of credits rather quickly given their new pricing.
Huh, I find my copilot plugin to be so incredibly glitchy. My agents are always reporting that their shells are mangled that commands are truncated and all kinds of nonsense. Sometimes they spin up dev servers fine other times it just hangs waiting for a terminal response. So far I have found relying on the CLI from the model providers to be significantly more reliable.
I do like the integrations with the IDE however, they are convenient for rapidly reviewing changes. I just need their terminals to actually work!
I had this problem and it turns out it was my oh-my-posh command prompt customization. VS Code injects certain control characters into the output stream for agents to observe events and the theming runs after those mechanics are hooked up so it can interfere. Updating to the latest oh-my-posh fixed it for me.
Here's the oh-my-posh GH issue[0] in case your problem is similar but not solvable with a simple package update.
Ive also had huge problems with copilot and terminals, mostly due to corporate antivirus which makes it impossible to install clink for example. So the Shell integration never really worked. It has been fixed for a couple of months though now, and its working, maybe not great but at least very good.
The only sad thing is trying to use tools in a VS developer prompt (and how could this not have been fixed ages ago, its literall YOUR OWN flagship product). It knows how to launch the .cmd for it, but thats incredibly slow for single commands. Would be nice if I could tell it to just use an open terminal.
You can pool credits through open router (afaik, I'm only using a single user account), but if you top-up $10 per user, per month, any unused credits will rollover.
Tbh I think it still works, but only because the new allowance will likely get used very quickly within a billing cycle - I'm expecting this change to increase our orgs bill significantly based on how many API credits with open router I consume in a weekend using a single agent in a pairing style.
The pooling will only be useful if you have a bunch of infrequent/low usage users that you still want to have licenses.
Which is almost guaranteed to be the case for a large org, considering everyone will want auto complete and PR reviews, but on average most will not be making a ton of agent use
Lots of us have noticed that usage limits for Claude have been nerfed in recent weeks/months.
If anything, these new multipliers are more transparent than anything OpenAI or Anthropic have communicated regarding actual costs and give us a more realistic understanding of what it's costing these providers.
The fact that we were able to get such a substantial amount of usage for $20/$100/$200 a month was never meant to last and to think otherwise was perhaps a bit naive.
This feels like a strategy from the ZIRP era of tech growth where companies burned investor capital and gave away their products and services for free (or subsidized them heavily) in order to prioritize user acquisition initially. Then once they'd gained enough traction and stickiness they'd then implement a monetization strategy to capitalize on said user base.
However, inference costs for entirely good enough models are likely to keep declining in the future. We're probably hitting diminishing returns on model size and training. The new generations aren't quantum leaps anymore, and newer generations of open source models like DeepSeek are likely to start getting good enough.
There's going to be a limit to how much they can raise prices, because someone can always build out a datacenter and fill it up with open source DeepSeek inference and undercut your prices by 10x while still making a very good ROI--and that's a business model right there. Right now I'm sure there's a lot of people who will protest that they couldn't do their jobs with lesser models, but as time goes on that will get less and less. Already right now the consumers who are using AI for writing presentations, cooking recipe generation and ELI5 answers for common things, aren't going to be missing much from a lesser model. That'll actually only start to get cheaper over time.
Also for business needs, as AI inference costs escalate there comes a point where businesses rediscover human intelligence again, and start hiring/training people to do more work to use lesser models--if that is more productive in the end than shelling out large amounts of cash for inference on the latest models. [Although given how much companies waste on AWS, there's a lot of tolerance for overspending in corporations...]
> because someone can always build out a datacenter and fill it up with open source DeepSeek inference and undercut your prices by 10x while still making a very good ROI-
Not sure how it all works out. Currently trillion dollar companies can't make a native app for platforms. Everything is just JS/Electron because economics does not work for them.
And here companies can make GW data center running very expensive GPUs for 1/10th of current prices. Sound little fanciful to me.
The price you pay for anthropic must include the price of training new and better models which is incredibly costly. If you use the models someone else already spend money to develop you don’t need to pay this price.
They've been like that for a while actually, I think at least since the big hype around ChatGPT 4.5 (or was it 5?) and that underwhelming, lukewarm, oversanitised presentation by Altman and his team.
Yups... Mythos is the smallest possible leap. Not a standard model generation advance, not even a version point advance. Just the smallest possible quanta of a change. We are absolutely hitting a plateau any day now. Any day. Any time. Any second now. Yup. Right now! Surely!
I mean let's be realistic - all that we know about the "mythical" Mythos is the carefully curated and release stuff by the Anthropic's PR team. Is it really a huge leap they are making it to be? I doubt it. In fact I bet if it was indeed that powerful and dangerous, as they imply, they'd find a way to release it immediately, devastate OpenAI and DeepSeek and secure a leading position in the market. Why is it not happening? I suspect because Dario is again at it, peddling his bullshit.
Yeah. AI progress is insanely fast if you compare it to anything else. Where else is a one year old technology already hopelessly outdated? 10 years ago is basically stone age.
I am continually tripped out by the fact when I was 16, I didn't have a 'smartphone' beyond a Windows Mobile 6 phone that had no internet on it.
Now, I have this high-resolution shiny object that can near instantaneously get any information I want along with _streaming HD video to it_ *anywhere*.
15 years even feels like a stone age. I can't fathom what it has to feel like people in their 60s and 70s.
I'm not quite 60, but it's always interesting to me that I feel quite the opposite of this. When I was 16, I didn't have a computer, didn't have a phone, had never used the Internet, but when I think of how life has changed, it's frankly not much. I woke up this morning, scooped my cats' litter boxes, took out some trash, made myself breakfast, ate that, read some news while eating, then lifted weights in my garage, had some work meetings, wrote up some instructions per a customer request from Friday, and am about to go drive to the lake to go do a 9 mile longboard loop.
That's very close to a normal day in 1996. The biggest difference is I read the news on my phone instead of a physical newspaper. The news was not any more interesting or informative because of that. I guess I can also still do the loop reasonably well, but I'm a lot slower than I was in 1996 when I was a cross-country state champion.
My parents are closing in on 70 and I guess I can't speak for them, but I'm at least aware of the daily routines of their lives, too. Walk the dog, do housework, DIY building projects, visit kids and grankids. Seems much the same, too, with the biggest difference being they're now teaching my sister's sons to play baseball rather than me, but shit, one of her sons even looks like exactly the same way I looked when I was 7! The more things change, the more they stay the same.
General agree... I still do the things (mid-50's) I used to do when I was a teenager with no computer, no phone.
But - now they are easier - I can read books on an e-ink screen and pretty much instantly find what I want to read next. I get my news on a phone. I used to watch TV/movies broadcast or on tape rentals. Now, I have just about everything I could ever want available - without ADs... those were such a time-waster.
What has changed is that I have access to MORE information than my local (or school) libraries could ever provide - in a variety of more accessible formats. Whatever tools I need to get "work done", I can find a myriad of free and open-source options.
But - the overall days and household family routines are the same - now, instead of reading a paper book while waiting to pickup my kids (or other family members) "back-in-the-day", I can read my device, or connect with my DIY communities online on my phone - or learn something new. I don't have to schedule life around major broadcast events, I can easily do many tasks while I am "out-and-about".
If your parents are closing in on 70, I would have expected you to be closer to not quite 50 than not quite 60.
I am just over 50 myself and I agree with your points. Technology has changed but life is largely very similar to wear it was in the 90s. At least day to day. Attitudes are way worse now.
I always wonder the views of older people. My parents are very technology forward and have been my entire life so it is difficult to gauge how different life is compared to when they were growing up.
It's easy to hear "Oh well I only had 640kb of memory and typed programs out of a magazine I got in the mail!" and see as distinct from having 'unlimited' resources and the internet.
Your insight is good ("The biggest difference is I read the news on my phone instead of a physical newspaper") that life sort of stays the same but the modality changes. People still go to the store like they did in the mid-1800s but now it is by car.
I wonder what our "industrial revolution" will be where the previous generation lived (ie: out in the country on a farm) totally different lives to the current (ie: in the city in a factory). Maybe when space travel and multi-planetary living is normalized?
> It's easy to hear "Oh well I only had 640kb of memory and typed programs out of a magazine I got in the mail!"
Since I was there (young, but there), I want to point out that this crosses three eras which all felt quite different:
1978: typed programs in from a magazine or loaded from a cassette (16kB, TRS-80)
1983: loaded programs from a floppy (64kB, Apple ][ and C64 etc)
1988: loaded programs from a hard disk (640kB, IBM PC and Mac).
Exact years vary but these eras were only about 5 years each. Nobody had a floppy in 1978 but almost computer user did by 1983; nobody had a hard drive in 1983 but almost everyone did by 1988.
To some degree this already happened with the move from the industrial city to suburbanization and then re-urbanization. In particular one of the most notable recent developments is that urban waterways are now pretty desirable places to be with parks and recreation; in most industrializing cities the waterfront was actively avoided because the industrial use made it polluted, smelly etc.
Depends on where you live. My dad is almost 80, grew up in a very rural area, and when he was 16 they'd just gotten indoor plumbing. Up until he was 14, his school was a one-room school house with no heating other than a wood stove. If you were the first kid to arrive for the day, it was your job to get the fire going in winter months.
Housework meant no laundry machine, no dishwasher, and possibly no vacuum cleaner. That means hand washing everything, and beating rugs with sticks and brushes to get the dust off of them.
And at some point even frontier model costs will hopefully come down (if there is still a meaningful difference between closed and open source models at that point) as all of the compute that's being built out right now comes online.
But the prices haven't been going up by multiples of 6 for the past few years. Things are actually changing now. I don't think it's over, but in the short term, it's going to be considerably more expensive.
They will smooth up the spike. Or be subtle and transform the existing quota so that they run out more quickly. Calling it caching, compression, optimisation, of course for the sacred benefit of the users.
Dunno, if in this day and age you are making inference more expensive, more scarce, you are honestly moving in the wrong direction and DeepSeek and others will gladly take your lunch.
> The hardware to run deepseek is still incredibly expensive.
Deepseek API pricing is very low compared to Anthropic/OpenAI API pricing.
For many, the 300% difference in pricing may be difficult to justify, if the quality difference is very small. And there will be many tasks where the most expensive/the best model, is not needed. Currently many people end up using Opus 4.7/GPT 5.5 for many tasks without thinking about it.
Near zero probability of that. The model is more efficient and the company who trained it did not blunder trillions of dollars to do so. China has better electricity infrastructure than the US too, so the likelihood they can scale out before the US ever could is high. Long term deepseek, Alibaba, etc hold the most cards for sustainable AI even despite the attempted Nvidia embargo
I am not shilling China, this is just what is happening right now.
I think the Chinese government works differently than the US government. I think China has been subsidizing their electricity grid for decades and leading the world on sustainable electricity namely solar. While the us has let their infrastructure rot and laughed at government inefficiencies for about half that time. The US has data centers running on gas right now while waging wars blowing up gas infrastructure world wide. It would be comical if it wasn't an environmental disaster. Most of them have no hopes at even getting enough power in well established areas short term.
I realize what I am saying may come off as propaganda because the US holds net negative views on China so here are some links.
I think because openai spent so much money upfront showing how it was possible to do this and laid out a product roadmap China got to get on board much cheaper and easier. I see no reason to not believe any of these companies when they say they didn't squander tons of money to do what they did because I don't know how openai has even spent all the money they have it's actually ridiculous to think about.
It's not really about that. China is eating the US's lunch when it comes to ai. Don't get me wrong opus is the strongest model out there today, but that's the us's only advantage right now. Deepseek,qwen,kimi, etc all have fundamental research making the models smaller, more efficient, scalable, etc. in the US the plan is to buy all the hardware, write legislature, embargo other countries, keep models and research closed, so people cannot innovate for the next two to five years.
Unlike the us chinas focus is on research and sustainable building. China also has really good infrastructure for energy, etc. it is also to their advantage to drop 5 billion instead of 2 trillion and beat the us while turning a profit.
Chinas focus in ai is less flashy and because they are the biggest manufacturing super power in the world right now, it directly feeds their economy. They aren't looking for applications or to replace thought workers with slop bots, they have natural needs for this technology. Us manufacturers can't compete so they have to keep companies from selling their goods there see byd. China sees it as commoditizing their complement, the us is risking its entire economy and it's environment and resources, kind of scary.
If/when it gets to the point where it can replace a skilled worker, the service can be sold for close to the same price as that skilled labour. But the AI can run 24/7, reliably, and scale up/down at a moments notice.
There's not going to be much competition to drive prices down, the barriers to entry are already huge. There'll likely to be one clear winner, becoming a near-monopoly, or maybe we'll get a duopoly at best.
Yes, a lot of people (not me). Why? Well because that was the whole value proposition of these companies, relentlessly pushed by their PR and most of the media- rememmber it was something something Pocket PhDs, massive unemployment etc?
"There's not going to be much competition to drive prices down, the barriers to entry are already huge. There'll likely to be one clear winner, becoming a near-monopoly, or maybe we'll get a duopoly at best."
Based on what exactly? So far every time OpenAI, Anthropic or whatever has released a new top performing model, competitors have caught up quickly. Open source models have greatly improved as well.
I expect AI to be just like cloud computing in general - AWS, Azure, GCP being the main providers, with dozens of smaller competitors offering similar services as well.
Right now China is flexing the future in my opinion. Smaller, widely available, frontier models for pennies on the dollar.
I think the future of ai will be breakthroughs that let it run on commodity hardware, and the average person will not be paying for it from the cloud unless they want to be surveilled or are stuck on older hardware.
Right now I am running about what was a frontier model 1-2 years ago on a junk machine. Some people are running what was a frontier model 4 months ago on PCs and laptops that cost 5,000. In a year I think the landscape will be even better.
I do. "Commoditize your complement". Want to sell lots of silicon? Give away good local models to run on that silicon.
Even if SOTA models in the cloud are a few percentage points better, most work can be routed to local models most of the time. That leaves the cloud providers fighting over the most computationally intensive tasks. In the long term, I think models are going to be local-first.
(Unless providers can figure out a network effect that local models can't replicate).
Why on earth would that happen when everything else is moving into the cloud to tie it to ever-escalating subscription fees and prevent piracy?
Even with gaming, where running high-end 3D games in the cloud seems like madness and inevitably degrades the quality of the experience, they won't stop trying.
> In the long term, I think models are going to be local-first.
Why? There's an inherent efficiency advantage to scale, while the only real advantage for local models (privacy/secrecy) hasn't proven convincing for broader IT either.
It's foolish not to care about privacy especially as a company. You know how it prevents you from emailing yourself your tax documents? Meanwhile thousands of employees are sending literal design docs, software, product goals, etc to several ai third partys. Not only is that insane, the companies they are sending it too intend too and openly admit to scanning the data, make software products themselves, and intend to create models that can produce their products automatically.
The reason local models hasn't caught on is several fold. It's marketing to say your company follows the latest trend, and there's an inherent pressure to keep AI companies afloat so the economy doesn't entirely collapse. The other is, it wasn't until the last month that these models have caught up to frontier models. They just did, and they are more efficient and don't require a team of 500 to deploy.
Local first models aren't just more private than the API vendors, they also have the advantages of fixed cost, lower latency, and better stability - local models don't get nerfed/"updated" in the background like chatgpt does.
Maybe in a world where these AI companies behaved with some semblance of ethics and user-friendliness they would be on even ground, but for anyone paying attention local models are obviously the future.
That's a silly reason. For non-agent use cases what kind of utilization are you going to average on your own GPU, 5-10%? And that's without batching.
Even with overhead and scaling for peak use and a large profit margin, any company with an ounce of competition will be vastly cheaper than self-hosting. And for models you can run yourself, there will be plenty of competition.
"This change aligns Copilot pricing with actual usage and is an important step toward a sustainable, reliable Copilot business and experience for all users."
I see statements like this as strong indicators that the sales people are wrapping up their work and the accountants are taking over. The land rush is switching to an operational efficiency play.
Just be aware OpenRouter charges a 5.5% fee, I didn’t know until recently. I like the product, and I think the fee is fair, but if you want the absolute best pricing then go direct.
But with open router you can always just use the latest model. If you're committed to eg Claude opus then you're better off going directly to anthropic for sure, but if not, varying other models may be fine too, depending on use case and be massively cheaper. Eg new deep seek model with same mio context window or Kimi k2.6 with 270k context window for subagents which implement
>but if not, varying other models may be fine too, depending on use case and be massively cheaper
Do inference providers have standardized endpoints, or at least endpoints compatible with claude code? Otherwise to pay 5.5% on all your tokens just so it's slightly easier to swap providers (ie. changing a few urls?)
There's nothing trivial about getting a Google API key. Openrouter removes that stress from my life. And I can route requests to providers above a certain TPS threshold. And much more.
I had copilot mainly so I could write issues and throw agents at it, while I went off and did other things. Has been great for contained spot work.
At this point, I'll go ahead and leave it expire, and then consolidate between Codex and JetBrains AI. Especially since Xcode supports Codex with a first-party integration.
They do for any new plan.
Those multipliers are only for people that paid annually. After their subscription ends they'll go into token based pricing like the rest of people.
I understand it like : the 10 usd is for handling the business record, maybe also the harness, I get a few coins to kick tires, but to use it for anything real it’s pay as you go by the tokens list price.
I thought I was smart for buying the annual plan after I graduated and lost my student plan and then GitHub taking away my Copilot Pro I got for free for being a author of a popular OSS project. Turns out I'm being punished for making that year commitment to them. I like to think I'm only a moderate user of GHCP so this is just terrible for me. I'm honestly thinking about cancelling and switching to alternatives while also looking at investing in a local LLM setup.
27x for Opus is genuinely shocking. at that point you're not paying for convenience anymore, you're just paying a GitHub tax. OpenRouter or direct API makes way more sense unless you're really glued to the IDE integration.
Does it effectively bypass regional restrictions for you, so you can use something like the Claude API from unsupported regions such as Hong Kong, or does it still enforce the official providers' geo-restrictions?
OpenRouter is great for budget control, but as they are indirect APIs, your experience with cached tokens may vary, eventually costing much more than in direct depending on the providers.
You can pay with crypto though, which seems to be convenient for people under sanctions or with limited access, or if you are in low-tax jurisdiction (e.g. HK)
That said I think few people using openrouter are actually being selective about providers.
It took half a day to get my opencode setup, was not friendly. A lot of manually cross referencing model and providers. I was actually mainly optimizing for relatively fast providers. It all is super fragile and I'm sure half out of date; I have no idea if these picks are still fast, no promises they are still the same price (pretty terrifying honestly).
I'm mostly on coding plans so it doesn't super affect me. But man is it a bother to maintain.
The point of this loss leading is to properly hoover up the money in the pockets of enterprise customers, get them locked into the idea that they need the latest and greatest cloud-based model, while simultaneously starving everyone of the memory they'd need in order to run competent models locally.
In not-too-distant future we're going to be running better models on our phones than we can buy access to today in the cloud. Skate where the puck is going: soak the customers until that day comes.
It's interesting that the cost multiplier for Claude Sonnet 4/4.5/4.6 varies so much (1/6/9), while the API cost is exactly the same for all three models.
Also, the multiplier of 27 for Claude Opus 4.6/4. is way higher than the increase in API price would suggest.
On GitHub copilot you pay per prompt. More powerful models can do a lot more work (consuming a lot more tokens) per prompt. Also, they tend to use more thinking tokens.
> More powerful models can do a lot more work (consuming a lot more tokens) per prompt.
That is not my experience. Each model since at least GPT-4 can fill up an entire context window.
In fact, more powerful models can solve tasks faster, so their ratio of multiplier to API price should decrease, not increase.
For example, Claude Sonnet 4.6 has a multiplier of 9 and an API price of $15, which is 0.6 multiplier per dollar.
Claude Opus 4.7 has an API price of $25, so it should have a multiplier of 25 * 0.6 = 15 when extrapolating from Sonnet, but the multiplier is 27.
> Also, they tend to use more thinking tokens.
That might be it. Is there any data on this somewhere?
Those multiplier are only for grandfathered Pro an Pro+ plans that had annual billing, basically a way to scare people of out of those plans.
Ant new ones (and bussiness+enterprise plans) will be on token based billing since June 1.
One theory of the play of SpaceX might do if everyone migrates to query-based billing:
Provide cheap and unlimited access to Grok for programmers (hence the Cursor partnership/purchase for distribution).
-> This would drag massive revenue right before the IPO announcement, like if the company is super growing
-> At a loss, but don't worry, we need these funds to build the biggest datacenter of the universe.
This announcement would create enough momentum to increase valuation, and because of the merge of his companies, would save his X/Twitter investors from a tragedy.
-> Would also be a great service to Cursor investors and so, who are stuck with their VSCode fork
Which in turn owns Twitter. SpaceX is now a social media company in addition to a rocket company.
One theory I think Matt Levine posited, is that SpaceX will go public with dual-class stock that gives Elon control even with a minority ownership stake, and will subsequently buy Tesla, which doesn't have dual class stock, making SpaceX the singular "Elon Musk company", with him having operational control despite being public.
Show HN timing matters more than people think. Monday-Thursday, 9-11am Pacific, is when the front page has the most engaged readers. Weekend posts get less competition but also less engagement.
The routing automatically routes you to other inference providers (for the same model) if/when the original provider goes down.
It's a convenience cost, for sure, but it's not valueless in a fast-moving world. Certainly if you're comfortable with one provider and it's cheaper, do that.
For me the largest value-add is the unified API. Being able to instantly start trialling a new model with zero code changes is well worth 5%. The other part is not having to deal with billing for multiple platforms.
What's annoying is that it's obvious. In the case of GPT 5.5, if Copilot is going to charge 7.5x what GPT 5.4 costs while OpenAI themselves via the API/Codex only charges 2x of what GPT 5.4 costs, that will immediately raise an eyebrow.
To anybody who's been watching the tech sector with a critical eye for pretty much any period from the late 90s and onward, this is just the enshittification process. For most of OpenAI's existence it's been obvious, to me, that investors were burning insane levels of capital to build the market, and now that folks are locked in, you're seeing higher fees, ads, etc. Yet again, the user is the product; the investors want to siphon your data, attention and once you're hooked, money. And for companies like Microsoft and Apple, those hooks can dig deep.
Let's call it for what it is dumping. Dumping things on market below cost of production. This should not have ever been allowed. RnD costs I can accept somehow. But in this case the interference should have always been billed for the real costs that it took to produce and pay off the capex.
If you paid attention to the power requirements and amount of hardware being put into data centers, you should have realized that it cost them an order of magnitude more than you were being charged. To rework your analogy: they hooked you, now they're gonna see if they can reel you in.
They can only reel you in if its worth it. I still can code.
And while i do not spend 200$ privat, in my startup we discussed this and our current mental model is, that instead of hiring someone new, we prefer to have more money for tokens.
This is easier for us and has a bigger benefit. The cost of a new / first employee is very high, a 200$ subscription is not. Upgrading that to lets say 400 or 800$ is still alot easier and if i can run multiply and better agents with that money, lets goooo.
I'm looking at education -- teachers and students, not terribly tech savvy, are being mandated to use these tools. And then comes the rug-pull. It was worth it, but now it's outside of their budget. Poorer schools / students can't stay at the cutting edge; richer schools / students can.
“Enshitification” is just when unsustainable subsidies end?
Another reason to hate that word.
From a different perspective, you were granted an incredible gift from the companies who let you use their product on their dime. Hopefully you made the most of it when you had the opportunity.
No, it's much more than that. It starts with unsustainable subsidies, as Uber undermined the taxi industry with a ludicrous burn rate. And then, once everybody's hooked to the point that they can't imagine life without the product, you raise costs. And you iterate: raising costs, lowering quality, selling data, increasing addictiveness. Until everybody wants to get rid of it, hates every aspect of it, but is still hooked to the core product. I'm personally not using these tools, not using uber or Meta products. But I'm still using some Google products and it's hard to extricate them from my life now that I'm using them.
OpenRouter doesn't even have hardware. What are they possibly subsidizing? The platform costs?
OpenRouter is guaranteed to be about the highest margin operator in the business right now. Everyone wishes they'd be them, skimming 5% off as the middleman without any OpEx.
> OpenRouter is guaranteed to be about the highest margin operator in the business right now. Everyone wishes they'd be them, skimming 5% off as the middleman without any OpEx.
The 5% fee probably has to factor in Stripe's fees, which would be around 3% to 4% depending on whether it's an international card.
Streaming, caching, and tool calling can get pretty expensive with scale, even when you don't touch inference. Maybe they're doing something clever and are quite profitable.. or maybe they've already taken $40mm from VCs and are currently trying to raise $120mm at a 1.3B evaluation.
They also show headline prices for the cheapest provider of whatever model, but then need to hit different backends some of which may be more expensive. For now they absorb those costs, but the VCs always come knocking.
Just my opinion though. Totally agreed that they have one of the best positions amongst all AI providers from a financial standpoint.
> They also show headline prices for the cheapest provider of whatever model, but then need to hit different backends some of which may be more expensive. For now they absorb those costs, [..]
They do?? I was under the impression I was just playing the price for whatever provider they deemed 'best' for each completion.
Checking now: The way they describe it in their FAQ is that if the price changes, then they will bill you the new price. But I read that as regarding if the primary model provider changes their headline token cost; not in the case of pricing differences for models that have many different backends that host them.
Regardless, I would be more concerned about the streaming costs if the service continues to blow up and they scale aggressively through VC investments. If their 5.5% skim accounted for what they needed, you'd think they could effectively grow organically..
I am and I see it as stopping the music at a party when you want everyone to go home without telling them to go home. There is also the offer to quit with prorated refund for the remaining time. I think I am going to take it.
I don't know if it's just me but copilot kind of sucks. I've been running local models with like 9b parameters and they are about as good if not better. Obviously there's no integrations or whatever and I get most people are probably paying for that than anything else but eh. Big no thanks from me.
That's so unfair to us hard working developers. A month ago i could buy for .4$ a turn with Sonnet. Now i have to pay at least .9$ for this turn. Weeks ago i could buy for .12$ an Opus turn after they already raised prices and now they want .27$ from me for the same product! They are stealing from us!
They aren't stealing from us, for several reasons. First of all, it's a voluntary transaction. If you don't like the prices, use something else. Or don't use AI at all.
Second, you have no idea what their costs are. It is most likely that they are simply passing on their costs to you. If that was not the setup, users would just go to another service provider who was providing tokens at a cheaper rate. It's not like there is a dearth of competitors in this business.
"Your plan pricing is unchanged: Copilot Pro remains $10/month and Pro+ remains $39/month, and each includes $10 and $39 in monthly AI Credits, respectively."
If there's no discount on credits (in terms of tokens per dollar) over other providers, I'm going to switch to a PAYG provider. If there's a month where there's little to no coding I can pocket the 10$. What incentive do they give to stay with this plan?
Or if you're a business with multiple seats, these plans may be more inefficient than raw API usage billing. Since if anyone at your organization fails to utilize their full $19/39 allotment each month, that's wasting money, whereas with API credits it is 100% utilized.
I don't think they've thought through the implications of this. Everyone should cancel and go usage-based billing with caps.
They do address this in the doc, Orgs can now (although it was vague as to whether it was an option or just the new standard, probably option due to business contracts) 'pool' the Usage billing across all users.
I'm guessing they did that (and the 'temporary bonus credits') to make the pill easier to swallow for that side of customers.
It forces you to pay at least $20 in tokens per user even for people who use less (they probably have stats on how many people use just autocomplete, which doesn’t count against the quota. or have a seat and don’t use the service at all).
You can get at least at baseline vague stats but from what I have seen it is more account than user focused, i.e. 'X number of users used %feature%' and/or '%model% was X percent of requests per day'.
That said it's worth noting, I don't see how anything they expose will reliably help orgs plan costing from what AFAIK is in fact a big shift for billing/costing planning.
> It forces you to pay at least $20
For better or worse that's public pricing, i.e. if you are coming in also negotiating VS for devs, windows/office licenses for the rest of the business and stuff like Azure Devops... a lot of their stuff gets cheaper if your company's IT procurement group is vaguely competent at negotiating. Not even talking bigcorp here I'm talking 500-1000 employee range.
Of course, very small orgs will suffer, but it does tie in with the theme over the last two weeks; anyone with a personal account is basically subsidizing the credits for the business accounts during the transition period.
This was my first thought too. "Oh cool, I should be seeing lower prices" as I don't use Co-pilot that often anymore. But no, that's not the case. It rather served to remind me that I should probably just cancel.
Seems like a strong signal the money burning party is coming to a close. Nearly all AI companies have tightened their belts in the past month. Anthropic removed Claude Code from the Pro plan, Z.AI increased their prices, GitHub removed some Claude models from Copilot, now this.
Also, Opus 4.7 seems like a model more intended to save Anthropic money than push the bar.
> Seems like a strong signal the money burning party is coming to a close.
One provider who was undercutting the market with non-standard billing model moving to a more standard billing and prices doesn't seem like that strong of a signal, other than that Copilot was underpriced.
It was the only clear model from a user's perspective. Sure, a request may not perform as expected, or end earlier than desired, but it was an agreed to cost that was clear on both sides: 1 enter press in a prompt window = 1 request.
If they wanted to limit what a request can do via their harness, I'm sure they artificially could.
I hate all of the other plans I've seen of here a "credit" or here's a "bucket of usage", and we pull an announced amount from it based on arbitrary info that can't be audited or proven, and most of shat is spent might be entirely useless anyway.
Claude Code has a problem where 1 request could take a significant portion of your 5 hour window, and it's unclear why.
It's much like SEO, where Google sometimes says things that might help, but it's just magic wand eaving hoping something works.
the point is that they tipped their hand about where they want to go in the future. They are just A B testing to see how much it pissed off their customers
I signed up over the weekend and still have access to Opus. I believe the AB test they were doing only removed Opus from a small percentage of users.
Don't think I'll be renewing though. The usage limits are low enough that I don't think this is worth it. One complex prompt while Americans are awake will wipe out your alloted tokens it seems.
Yeah, honestly it feels like this came faster than I was expecting. I thought we'd see another few years of reeling in with too-good-to-be-true prices to really lock in dependency but it feels like most companies have kind of a lot of wiggle room to back out of this still
Not really sure why I would stick with Copilot after this, and increasing Sonnet from 1x to 9x for annual subscribers is highway fucking robbery. Very glad I didn't commit myself to an annual plan.
> Alternatively, they may convert to a monthly paid plan before their annual plan expires, and we will provide prorated credits for the remaining value of their annual plan.
I don’t understand if this means they’re providing actual refunds or not. For them to straight up go back on their word this had to have been a major cost they didn’t exactly expect.
Save us Deepseek!
I don’t need the world’s greatest programmer for the types of vibe coding projects I actually build.
However, if compute keeps going up in cost, hiring skilled people who know how to utilize it becomes more important. This might save the tech economy.
Contrary to the other reply, I'm going to say yes, that's exactly what it means. For Github Copilot users with annual plans that are grandfathered in to per-prompt rather than per-token pricing, Github is increasing the cost of Sonnet from 1 "premium request" per Sonnet prompt to 9, thus meaning that those users will be able to submit 1/9th the number of prompts per month before incurring additional usage charges. For all practical purposes, this is a straightforward 9x increase in price.
Not quite. Premium models have different type of multipliers applied. The multiplier decides how many PRUs (premium request units or tokens) are used. These PRUs are replaced with different units with this announcement but the methodology remains the same: https://docs.github.com/en/copilot/concepts/billing/copilot-...
I'm starting to see comments like this in a new light after using some primarily AI-coded apps the past few weeks. They are a lot like apps that were built by hundreds of developers/product people over years and years, in the worst ways.
Inconsistent design patterns from page to page, half baked features, inconsistent documentation (but BOY is there ever a lot of it!), NIH ui component libraries that don't act like you'd expect. All that fun stuff.
It's like they speedran the worst parts of enterprise apps.
I made so much progress on my personal projects, I actually regret not subscribing sooner. I've been coding alone for over a decade. It's been great having a coding buddy for a change. I'm actually going to miss it.
(This was originally posted to Microsoft and OpenAI end their exclusive and revenue-sharing deal - https://news.ycombinator.com/item?id=47921248, but in a perhaps-futile effort to keep the discussions partitioned, Maxwell's demon will move it to the Copilot pricing thread.)
Everybody who says it's a 5-9-27x seems to not be aware of the obvious loophole. More like 50x increase. You were able to use over $500 worth of Opus on a $10/mo Github plan easily, no hacks. You could just prompt "plan this out for me, don't stop until fully planned, don't ask any questions", and you would get ~$5 worth of planning in one 3x request. At 100 requests/mo, each easily reaching $5, that's easy $500 worth of tokens.
It was going to happen regardless due to the nature of enshittification. If they really wanted to stop people using 100M tokens a day, they could've prevented it years ago.
So, silicon valley decides to use their playbook of expand at all costs by burning money to acquire the market (like a carcinoma), and it is the users fault ?
Should we be blamed about uber destroying the taxi business, or airbnb the hotel one? Oh sorry, "disrupting".
Uber was dirt cheap, now it is the same price as taxis, and the people working for it (the "partners", not employees) have no social benefits.
Airbnb was cheap and humane, now it is THE cause for housing crises and massive residential property "investment".
The playbook of silicon valley is destructive, not disruptive.
It is by design aimed towards wealth accumulation. The ones with most money can capture the market, and make even more. It really is late stage capitalism.
And the more wealth inequality there is, the more pain, poverty and instability will be as well. AI will only exacerbate this.
Yeah it was crazy. Nowadays I use pi with OpenAI GPT 5.4/5.5, which to me seems both better and more generous than Claude. I supplement it with OpenCode Zen to get access to a bunch of models at token cost, and OpenCode Go ($10/mo) to get subscription-style access to Kimi, GLM and friends.
That was not my experience. When I tried to use Opus for longer tasks with Copilot, it would fill up the context completely and then crash without any output, while still consuming premium requests. (At least from September 2025 to January this year. Haven't tried after that.)
Copilot has improved immensely in 2026. I'd say to give it a try again if you're up for it. It works about as well as Claude Code these days in my experience.
Unfortunately, Opus was removed from the student plan in March. So far, I had been happy with GPT-5.3-Codex, but that model seems to have been removed this morning.
When using opencode or copilot CLI, the error messages are displayed
normally and it's possible to see what's going on. Under Pi, it
sometimes just hangs, or Pi crashes with some bun stacktrace and that's
it.
Copilot has introduced additional limits for Claude models in past
month, and it's rather easy to hit it. Pi often doesn't show anything
when this limit hits (although sometimes it shows the error, I guess it
depends on Pi version).
This was my solution to very very though compiler tests that would take sometime up to 4 hours to figure out. Some of the time would be spent on running the tests, but still... I was burning so much tokens. I have free Copilot for my open source work so I wasn't even paying the $20.
The 27x multiplier on Opus is the tell. That is not a pricing model designed for broad adoption, it's a price signal that says 'use the cheaper model.' The problem is that once users start self-censoring which model they reach for based on cost anxiety, you've degraded the product experience in a way that's invisible in the metrics but very visible in churn.
Flat subscriptions had one big advantage: zero cognitive overhead per request. That's worth more than people admit.
This may be a more accurate analogy... "The Porsche you rented at $200/mo now only allows you a maximum of 100km of travel. You will be automatically charged extra when you go over that."
On top of being worth less, the subscriber discounts are gone.
The old plans were $0.033/request for Pro, $0.026/request for Pro+ and $0.04/request for pay-as-you-go. That discount is now gone. They even still advertise "5x the number of requests" for Pro+ over Pro.
Yeah, if I go to a petrol station with 50€, but only get a tenth of the amount of petrol I got last week, I may think that the price has in fact changed.
It’s technically true that the plan prices haven’t changed, it’s just the value you get from those plans has plummeted. It’s classic deceptive sales language.
I was curious why a company would still use the VS Code + Copilot sidebar method for coding, rather than something like Claude Code. Turns out there’s a GitHub Copilot CLI!
I thought I was pretty familiar with available options, but no one in my circles ever mentions this product. It doesn’t seem to have much mindshare.
I'm curious about the opposite: Why would anyone use the CLI when, at least with Copilot, the VSCode plugin is super tightly integrated with VSCode, meaning the agent can see everything I can see. There's no mismatch in linter calls where I can see a lint in the ide that the agent can't find for example. I've had this problem even using CC in their VSCode extension, so I can't imagine it's not an issue in the CLI as well.
I use the Claude Code VSCode plugin for 80% of my work.
I prefer it because I can look at the code (although not as often anymore) and config (very often!) easily.
It also lets me jump to previous conversations easily.
There are a few cases where the CLI makes sense. One big one is if you are running multiple simultaneous sessions on a remote server using Tmux to have them preconfigured when you reconnect is nice.
The vs code integration is pretty slick. I can copy and paste function names into the prompt and it automatically turns them into these `#sym:` reference objects that I presume populate the context window with metadata about the function and where it lives. It knows what file I'm currently looking at as I jump around in the code, and that automatically gets loaded into the context. I can also drag and drop folders or specific files for context into the sidebar.
It's a lot of stuff that makes me have to type less into the prompt, since it's already getting so much info from my editor
I’m actually trying to move back from the Claude Code style, I feel like it’s easy to become distant from your own code, and I am feeling uncomfortable with that.
I’ve “vibe-coded” some projects and when I start to find issues or go to refactor them I don’t have that memory of why decisions were made, because many decisions were never made.
I've used it quite a bit. There are a lot of AI terminal coding products and this is another one. It works well, handles sub-agents without issue and does a reasonable job operating in the Copilot ecosystem. It handles mid-task questions and such we well.
But its a really good UI for agentic coding. Not sure why more people don't use it. I've tried the others and keep coming back to Copilot chat. It's a really good tool. Which is why the rugpull on pricing is so concerning.
Yeah, I've been using it heavily at work since the beginning of January (and have a personal Anthropic sub to compare to). Copilot CLI is pretty good, honestly. Most new features in Claude Code get cloned by Copilot CLI within a couple weeks. Claude models seem mildly more clumsy in that harness than the one they're trained on - subjective guess around 20% more turns for an equivalent task - but it's not a noticeable difference in the final output.
The other cool thing is Copilot SDK, so you can build agentic capabilities into apps, or build tools, that leverage the agent harness of the Copilot CLI:
> I was curious why a company would still use the VS Code + Copilot sidebar method for coding, rather than something like Claude Code.
I use Claude Code, but I kept my Copilot subscription around mostly for really cheap usage of other models when I need to try a different one (which appears to be ending, in a sense) and also the autocomplete in Visual Studio Code which was really great across a bunch of files, I could make changes in one file and then just tab through some others.
I wonder what other good autocomplete is out there.
> also the autocomplete in Visual Studio Code which was really great across a bunch of files <...> I wonder what other good autocomplete is out there.
I am in the same boat. I tried looking for tab/auto-complete implementations ~ a year ago and it was pretty disappointing. If that has changed, would love to know!
Windsurf has free unlimited tab complete, I use it as an IDE alongside Claude Code and it works pretty well. I think Google's Antigravity also has free tab complete but no idea if its any good.
That's espoused as the big reason for the price increase: most Copilot subscribed developers it seems have moved to "agentic usage" with the CLI and Cloud-based agents.
Which feels a bit like a kick in the pants for me as a developer that was primarily using Copilot for VS Code ghost text and very rarely used the Chat sidebar much less "agentic" tools.
Copilot Pro sort of made sense for my personal account when amortized across a year, but I don't want to "waste" $10/month on credits I won't use most months.
Liability. I work in a tightly regulated market and I cannot use any service due to be in touch with PII data. If it was my option I would just use OpenCode with deepseek, it is more than enough for the majority of my tasks.
I'm just so confused why people aren't just using ghostty/kitty/terminal.app and claude code. Compared to the other approaches I've tried, it's by far the most effective way to get performance from opus 4.6/4.7
I don't know about others, but I use Copilot more often than other apps because of its tight integration with the VS Code itself where I still spend most of my time working on other things while letting AI do some task that I decided to delegate to it.
I tried the VS Code + Copilot sidebar approach a few months ago. It was definitely rough around the edges compared to Cursor/Claude. In our corporate environment, we weren't even able to use frontier models.
I've been using it for a few months because Copilot was the only AI blessed by our corporate overlords. It's not bad, I would say it's about 80% as capable as Claude Code, which I've used extensively on personal projects. However CC was recently approved, and I'm betting that with these changes to Copilot pricing we'll end up dropping it like a hot potato.
Search has become so bad that I also struggled to find Claude Code alternative and made my own tight (not editors, not plugins, not agents, strictly similar to Claude Code CLI) list: https://github.com/omarabid/cli-llm-coding
The list is not long but there are quite a few options. Even Grok has its own CLI!
The reality is, even though a CLI prompt looks very simple, it's a very complex piece of software. I personally use Claude Code (with GLM) and anything else I have tried was significantly inferior (with the exception of opencode).
I don't use Copilot or any paid AI but all of this usage-based billing reminds me of cellphones back when you paid per individual text message.
Usage paying for AI is 1000x crazier because you're not even getting a guarantee in the thing you pay for in the end. You have to keep feeding it prompts and hope it gives you the solution you want. You may end up with no expected result yet you are paying for it. At least with texting, you got what you paid for.
I wonder how long it'll be before all AI costs are flat unlimited monthly fees or even free across the board, without compromise.
I expect in the future we'll find out that someone in the industry was juicing the numbers with fake thinking tokens or something. The whole pricing model of charging you for the tokens it generates while not knowing how much it is going to generate going in has always been pretty crazy.
It reminds of early smart phones when the cell providers pulled away from unlimited data...and then they brought it back in s few years.
I think competition will get fierce. We see many people are attracted to the price stability of GHCP - it became clear what a request could do - the problem is that they didn't match results with cost. It's not clear what a 5 hour usage window in Claude Code can do.
There's no reason the harness couldn't provide a quote on the next request, aside from it takes effort and it would be upfront to the user, creating expectations.
Yeah, this was my frustration with Suno and Sora. You can burn a lot of credits (not to mention time) generating things that aren't what you wanted.
I don't mind a PAYG model for a simple chat interface. But when it comes to actually producing things, you burn through TONS of tokens creating the wrong output.
It incentivizes you to do most of that prompting on your own hardware/time, and only feed the final prompt with only necessary context to the big AI in the sky. It might even force you to think about the problems yourself for a bit!
> I wonder how long it'll be before all AI costs are flat unlimited monthly fees or even free across the board, without compromise.
That's already the case if you can self-host an LLM; you don't even need a mythical H200: gamer-grade GeForce cards can get you a long way there (if this page is to be believed: https://www.runpod.io/gpu-compare/rtx-5090-vs-h200 )
...after RAM prices return to normalcy, of course - and then wait another 2 or 3 generations of GPU development for a 96GB HBM card to hit the streets - and also assuming SotA or cloud-only LLMs don't experience lifestyle-inflation, but I assume they must, because OpenAI/Anthropic/Etc's business-model depends on people paying them to access them, so it's in their interests to make it as difficult as possible to run them locally.
That page compares models that easily fit inside the ram on either GPU. The biggest difference comes when one card can fit a model and the other cannot.
> In March 2026, Windsurf replaced the credit-based system with a quota-based usage system. Instead of buying and spending credits, your plan now includes a daily and weekly usage allowance that refreshes automatically.
With hindsight, per-request pricing makes no sense at all if an agent can burn a widely varying amount of tokens satisfying that request. These pricing plans were designed before coding agents changed the dynamics of token usage.
I wouldn't call it hindsight - I don't think anyone, at any stage, thought running a 10 minute+ sonnet session for 1 premium credit was ever profitable. We all knew it was a loss leader to get people using it.
It would have been profitable if that premium credit cost more than a negotiated discounted rate with Anthropic. We have no way of knowing if there were negotiated rates though!
There is no way to make that cost model profitable consistently. If 1 prompt can mean 100's/1000's of requests over hours, and you only pay for that 1 premium prompt, that can never be profitable.
They can engineer the harness to limit the amount it does. When pressing enter, it's be nice to have a "budget" per prompt, much like the model multiplier. When the harness used up the budget, it cleans up and cuts off the work.
But that would entail actual work and effort...and care for user's time and money.
Guys, you're discussing a house of cards to begin with: No matter how you're paying for the $CURRENTSOTA you're not garunteed that next month what you pay for will be the same.
So, lets do some honest evaluations:
1. The model itself is a non-deterministic engine of work with an unknown value; it's real value is just magic.
2. The business model itself is non-deterministic engine of profit with a known value; whatever the VCs have put into it, _must_ be piulled out. If Ed Zitron's numbers are correct, circa 2030, it's several trillion dollars.
So do some matrix multiplication of non-determinism vs determinism, and realize that the value proposition for _you_ is only going to decrease because #1 can never outpace #2, ensuring enshittification captures a smaller and smaller whale.
We know this. This has been the last 2 decades of money extraction from software. It was ok when it was some 12 year old's parents CC. But now it's you, or your business, that's going to either ben squeeze for value or squeeze out of the market.
And everyones squabbling about the color of the cost.
ok
The problem with assuming that tokens can only get more expensive is that the Chinese open weight LLM firms have dropped models which have a known, fixed price that can never get more expensive (since we can run them on hardware we own).
Well, I guess we're not discussing the same thing. The cost of cloud tokens are going to go up. They won't ever be cheaper. They're generating far more tokens than my AMD 395+ w/128GB at a much cheaper rate.
I agree though, it can't get cheaper than the cost of hardware it's just without sufficient documentation of the actual costs to run the cloud models, we can't really know what the "true" cost of each token is. I assume there's an economist out there somewhere that could figure it out though. Certainly, the cost should approach at a minimum a open weights model running on a local machine.
I've succesffully got Qwen3-coder-next to loop and generate sufficiently competent code and from what I can tell, the difference between this and the cloth is how quickly the gen happens and perhas how interactive it has to be.
I wonder if GitHub (Microsoft) is implicitly betting that enterprise demand is sticky enough to absorb these rates, especially given that Opus 4.6 “fast” was being listed at a 27x multiplier. Maybe they saw enough usage at that price point to conclude the demand is real. Or maybe the strategy is to keep the enterprise customers who can justify it while shedding heavier individual and power-user usage.
The interesting question is how long it takes enterprises to notice the capability/pricing tradeoff, and whether they respond by limiting access to the strongest models internally.
The part that worries me is that this market is still very early. Most developers and organizations are still learning how to use these tools effectively. Raising the experimentation cost this much may slow down the discovery process that makes the tools valuable in the first place.
As someone who works in a Microsoft shop with Copilot, I am curious to see what happens.
Due to data governance it will be difficult to move to a different provider.
At the same time, this price hike is so large that the ROI on copilot will be a net negative.
I think what will ultimately happen is that we will not pay Microsoft more than we currently do and we'll simply end up with less AI usage in the company and a reduction in productivity.
I also suspect that there are many "slow-moving", Microsoft heavy enterprises but with in-house devs that can't get anything but Copilot approved, and Microsoft trusts this will remain so.
It's not turning consumption based because there are a ton of these licenses just sitting idle.
As a single data point, this is absolutely true. At my current "Big Corp", Copilot was immediately approved while Claude is entering month 2 or 3 of trying to get approval.
Additionally, we got copilot for every user, including those that never write code or use AI tools.
yes. even in my province that's relatively behind on tech, at least +-20% were already strongly relying on Microsoft (Azure) and thought adopting Copilot was a no-brainer. That ofcourse comes with autonomous programming, because we're paying for it anyway!
These companies already lost the spark/innovation years ago. They're just using LLM's as a way to survive. Wonder how long that lasts.
As someone that is on the enterprise side in a non-tech F500 company, what I'm seeing is some FOMO and need to be part of the hype cycle. We're about to plonk a bunch of money on more Copilot licenses. Something got in the water where all the C-levels the past two months are pushing everyone to use AI but when they bring up examples of their uses its like "I use it to rewrite my emails" or prompt 'engineering' ideas that point more to patching over poor processes, data management, and decision-making within the organization or not.
What we're seeing across the board is every software company tossing AI onto their name or sales pitch and no one understanding what that actually means. But we will spend money on it because of FOMO.
I really question if we're reaching the end of the hype cycle to the point. I wish I were brave enough to put money on it. It feels like there was a command from up top to 'do something with AI' and leadership is scambling for some resume-building projects vs doing the hard work they should've done the past two years at a people and process level.
I'm at an org, where every medium to large company meeting talks about AI in similar breathless non-specific terms, yet m365 copilot is the only approved AI chat, and GH copilot the only AI coding tool.
Subsidies stop when LLMs improvement plateaued (though they still benchmark higher somehow). At some point, you have to make money or at least break even; and I think they concluded that we reached that point.
I really hope that they think so and that they're wrong and they get burned hard. Them and all the AI labs that lied, stole, inflated, hoarded and tried to justify all this as an existential moment where AGI would radically change society. I hope their calculus to reel in paying users is all wrong and now they all crash and burn instead of recouping VC money.
Github had, by far, the most easily game-able agent usage policy. People would force the agent to run a script before the end of turns that consisted entirely of `input("prompt: ")` so that you could essentially talk endlessly to an agent for the price of a turn. I see this less about the future of this industry and more about fighting the costs incurred by bad actors.
I never played any games like that, but simply giving the agent a clear exit criteria and instructions to check the exit criteria every time it thinks it's done on a complex task was often enough to keep it chugging away for most of a day on a single prompt in my experience. Per-prompt pricing just isn't sustainable period, even if everyone is acting in good faith.
I once asked it to do a comprehensive security review of our code. It churned for nearly an hour (and then produced 90% false positives). Insane that that usage was charged the same amount as me just saying "Hello".
After 2 months of using copilot pro, they've charged us 22$ in premium requests for my user, i spend roughly 1900$ in tokens, going of the on-demand pricing. This is estimated to be around 40-60% of real costs. They are undercharging by a factor 50 to 90!
One reason I used it was that I wasn't locked into a single provider and switching them was as easy as changing a drop-down. Small feature? Sonnet or GPT5.4/mini? Large changes? Opus. And why not see how good Raptor Mini does this one refactor?
It also helped build an intuition of what wach model could do and which parts it was weaker at because you could try them almost side by side, especially if one model's output wasn't great.
That said, these were all side projects so nothing truly consequential. Otoh, you might leave some extra perf on the table but I found the models worked quite with the Copilot harness.
Yeah, this is a very useful abstraction layer. The entire concept of separating the model creator from the model runner is good for competition and is customer friendly. Which means they likely hate the concept and want to kill it.
Gosh, imagine getting to do that with your TV/Streaming subscription. Getting to pay one fee to access some set number of hours per month from any of the providers.
The problem is I can't afford the tokens! Even on my $10/mo plan, running either 100 opus, or 300 sonnet agent runs would cost hundreds of dollars - well above my budget!
I liked copilot because I didn't have to think about tokens. I get hung up when having to think about the price of things, and its hard to think about the project at the same time I got to think about token usage like a gas bill. The usage system had its own issues, but having a set amount of requests was a very comfortable way to use a paid AI service.
Not paying per token? Not sending my code to someone else's servers for inference? That's the stuff of sweet dreams for a stingy, paranoid solopreneur like me.
If I could run a local model comparable to even Sonnet 4.6 without shelling out $50K in hardware, I'd do it in a heartbeat. But all I have is a 32 GB of RAM and an old RTX 4080.
Or am I not up to speed? Are there decent coding models that can run on dev laptops? Not that that's what you were suggesting by recommending a local model, necessarily; just curious.
I am trying to figure this out too... what I am seeing is that the local models like Qwen 3.5 family that fit on hardware like yours handle ambiguity poorly. But are capable of emitting complete apps too.
I do love using local models when I can, but qwen-35B is the best model I can run, and while its an insanely good local model, it does not compare to the big ones.
So given that I primarily interact with LLM's through VSCode, and I prefer the Copilot interface to the Claude Code plugin, does anyone have any suggestions on other plugins I should try? In my experience, Copilot is much more "plugged in" than any of the other plugins, in the sense that it can see things like linter outputs in VSCode. Basically, copilot "sees what I see" in a way that no other plugin or command line tool can, which make it much more ergonomic to use.
With this pricing change, I see no reason at all to stick with Copilot in principle, but I really need to solve this issue of IDE integration to move on.
You can use Copilot Chat* with basically any API provider, and if you switch to the VS Code Insiders build you can configure it to use literally any OpenAI API-compatible endpoint.
Other than that Zed has a similar experience which is pretty decent.
* By which I mean the good one, whatever it's called now - the part of Copilot that used to be a plugin and is now part of VS Code, not the thing that has always been part of VS Code.
Simply the output of the code written. It initially worried me that it was harder to add context (can't select a single line of code in VS Code and automatically attach it, for example), but I've found that the Claude CLI's output is so far superior that it doesn't even need that VS Code context.
Cancelling. Going with Codex $100, Kimi annual plan, DeepSeek API, and a local LLM once I get a Mac Studio.
Inference economics are going to be brutal in 2026 H2 when DeepSeek's new infra and model improvements come online, and Kimi launches K3. By brutal, I mean for OpenAI and Anthropic.
Yes there's no Opus at all on Pro. GPT 5.5 is also missing. Then again what would you expect, the economic reality is beginning to hit. Also I can't be too mad when the "base" models (GPT 5.4...) are still available and decent.
When I see how fast Codex max thinking GPT 5.5 eats our enterprise seat credits almost anything else seems cheap (until we switch our live systems from 5.4 api to 5.5 api I guess)... good thing I'm not the one paying for those credits and tokens (which is probably how most of the money is going to be made on AI going forward, borderline free chatbots for normies are done)
On my personal account, Copilot Pro+ still only gave me back Opus 4.7, whereas my work's Pro account still lets me use Opus 4.6.
So, my gut says, it's entirely possible that Pro+ will continue to have more segregation on model availability...
FTA
> Last week, we also rolled out temporary changes to Copilot Individual plans, including Free, Pro, Pro+, and Student, and paused self-serve Copilot Business plan purchases. These were reliability and performance measures as we prepare for the broader transition to usage-based billing. We will loosen usage limits once usage-based billing is in effect.
There's enough weasel wording here that I would expect only certain models get re-enabled on Pro.
e.x. lots of people seem to get good enough results from Opus 4.6, personally I prefer it over 4.7 in GH Copilot... locking that down to Pro+ would be, given this salvo of enshittification, a 'logical' move on their part.
You should indeed feel vindicated. Me too, I've been a paid subscriber for months and eagerly read every word. <wink> But yes, I feel like this is the first of several vindications you have coming.
What's the current situation for coding with Local LLM's on decent hardware? I have an M3 Max with 64 gb of ram and am thinking I should start looking at Ollama and Opencode? Is this a useful stack for smaller personal projects?
It’s getting there. You could give a try with qwen 3.6. It’s worth paying for better models in the cloud, but local models are now better than nothing.
One nice development recently was ollama's support for MLX optimization on Mac hardware. It's not obvious how to know you're using a model that works with it, yet, so it's rough around the edges.
I think that only applies to held-over users on the annual plan:
> Users on annual Pro or Pro+ plans will remain on their existing plan with premium request-based pricing until their plan expires, however, model multipliers will increase on June 1 (see table).
It isn't just the big multiplier increase, they also say "...and no new models or features will be added to annual plans going forward."
Can you imagine ten months from now and you're still rolling Sonnet 4.6?
Cancel/refund is looking pretty good. They're doing refunds until May 20.
"To request a refund, go to Settings → Billing and licensing → Licensing, select Manage subscription, then choose Cancel and refund "subscription". (The phrasing varies slightly depending on your subscription ). This option will be available until May 20."
It is an apples for apples comparison since those new multipliers only count if you are on an annual plan in which case the premium request system stays in place until you either cancel and get a refund or until your renewal comes up. https://docs.github.com/en/copilot/concepts/billing/usage-ba...
Those multipliers will only apply if you are currently on an annual subscription (and only until your renewal comes up or you cancel). So I assume they simply want to make it as unattractive as possible to get most people to cancel it and move to the token based system.
That's not an answer. It's specifically a discrepancy between 5.4 and 5.4-mini. If you look at all other models/generations you see that the cheaper model indeed has a lower multiplier. It's very strange that only 5.4 doesn't have this.
They're not the only ones in the AI sphere to wind back, but they're the weirdest case in my eyes. Microsoft invests in having engineers building open models and they don't use a single one. I really don't get it.
But what really surprised me most about Copilot is that it would bill you per question, nothing about tokens. So if I managed to produce a prompt that gave me back an insane amount of tokens for something, which using any Claude model would easily accomplish, you were giving me my money's worth, at your own expense. The math is not gonna math out forever.
One of the largest employers publicly engaging in a project which has the outcome of depressing wages. It's easier to "get" if you don't take the trillion dollar gorilla at face value.
The cheapest copilot plan felt totally unsustainable to me. For around £8 month i was getting 100 opus 4.6 prompts (albeit with a reduced context window size around 128k iirc vs 200k to 1m for first party hosted opus). Gpt5.4 was hosted with 400k context iirc.
On top of that, you’ve got 2000minutes of container runtime, so running cloud agents was included. As was anthropic agent sdk mode via copilot which is very comparable with claude code - not identical, the anthropic “modular prompt” is much leaner in the sdk version.
I cant say im mad, i got above what i paid in value. That said, going forward ill probably go back to openrouter payg rather than a subscription.
I got a free 3months of the gemini £19 plan and ive been playing quite a bit, 3.1 pro is a good model, i just find it slow. Flash i think i under appreciated until now.
I AM mad, because I just signed up for my private copilot sub. Otoh I do have it at work. I suspect we will start noticing differences there, too, but it all depends on the top users, vs average user since it is now pooled, IF they dont just allow extra usage. Bu that has become more expensive now, so not sure what they will do. I am hopeful that very few actually even use it. In my and surrounding teams, only maybe 1 out of 3 people really use it much.
I pay for Copilot annually, and mostly for its code auto completion features. I use CC if I want to do anything agentic. Not sure if I want to pay more for occasionally-good-intellisense at this point.
But you can no longer amortize annually, which makes it even more a question of "is this worth it this month?" each month. Especially for personal accounts.
I'm similarly thinking about sticking with the auto-downgrade back to Copilot Free when the annual sub ends and then just yelling about it any months I hit the 2000 completion cap.
I wouldn't mind a plan between Free and Pro that is just "all I care about is code completion and next edit suggestions".
I am seriously considering binge buying local AI inference hardware. The way this is going, there will be another big GPU crunch soon because everyone will need local models and/or open model inference capacity to do their programming tasks when the subsidized subscriptions are no longer flowing.
I doubt you can force them to provide the service with the original terms, but you might be able to ask for a (partial) refund. If not today, after a week of verbal abuse they will receive for this online.
It depends where you’re located. In the EU they have to honor the contract you entered, but presumably there is a clause that they can prematurely terminate the contract without cause and give you all of your money back (from the start of the contract).
In places with reasonable consumer protections (Australia, Germany) it almost certainly is illegal unless they give a full (whole year) refund. I think the short time limit of applying for a refund won't be looked at favorably either. Regardless of their ToS which I'm sure covers this.
But companies do lots of illegal things, and in general nobody takes them to court over it.
My thought exactly! First the usage limits + model limitations and now fundamental change to the billing. Hope some consumer watchdogs are looking into this!
That kind of clause would be void in many places around the world.
For example, the German Civil Code states:
Section 308 - Prohibited clauses with the possibility of valuation
In standard business terms, the following in particular are ineffective:
[...]
4. (Reservation of the right to modify) the agreement of a right of the user [TL note: this means beneficiary of the terms, eg. party or other subject of the contract] to modify the performance promised or deviate from it, unless the agreement of the modification or deviation reasonably can be expected of the other party to the contract when the interests of the user are taken into account;
"It" being the end of subsidization of tokens and plans (expected) but while lock-in to foundational models and cloud services is still lacking. Guess investors want their ROI sooner than later, given how big of a wrench the AI boom has thrown into global economics.
It's almost impossible to afford the usage-based pricing as a heavy user with multiple max and pro subscriptions. Seems like a bet that same-capability inference will keep getting cheaper on something like a Moore's law curve.
“Base plan pricing isn’t changing” is technically true, but for anyone using the more capable models heavily, this is still a price increase in all the ways that matter. The old abstraction was hiding compute costs; the new one mostly stops pretending.
Anyone paying enterprise billing can you shed some light how on the earth you will be able to fund this bill going forward?
Were you able to see assisted AI coding savings proportional to costs increase now you are going to get?
Companies removed people as AI assisted coding will be cheaper and now coding cost are going up from fixed $X to non-deterministic. The posts by Uber few days back about spending 12 months' worth of money in 4 months tells a lot.
Only path forward seems using Open-source models and many companies don't use Chinese that makes only Mistral one as the option.
Surely one path is to return to humans. We can accept that a profit grab to reduce staff numbers to maximise profits, maybe was never the best plan without being in control of the development tool.
Sample of 1 opinion: Massive downgrade for fun vibecoding on the go. Got done much before the recent rate limits on mobile on fun projects, their product went from fun to bad within a few weeks. And now into oblivion I suppose.
That was the first thing I turned off in VSCode. Autocomplete for my TypeScript projects was great. And the "AI" suggestions/completions were really getting in the way of me still being the "driver."
some of Github's open source maintainers have lost their free github copilot pro, guess this is really the next step for them to save cost in their infrastructure.
I started to use github copilot with vscode, but have never been too happy about the system. Over the months I gravitated to much more agentic workstyle, hardly ever editing much code by hand. The vscode IDE was getting more in the way. I had already started to look at OpenCode, and when I found it has a web interface, I was happy to switch over. I use a simple editor (KDE's Kate), or just less to skim through the code and/or a git diff.
OpenCode has some free models in it, but I think I will need to get some kind of subscription for a better one. But it won't be copilot any more. The market is moving so fast that I don't know what are the most resonable models, or the most flexible way to set them up so I can switch when prices change yet again.
Annual Pro+ subscriber here, mostly using it as a fallback when my Claude Max plan hits limits. The request-based pricing was genuinely the appeal, I could spill over from Max into Copilot without thinking about token costs. Going from 3x to 27x on Opus on annual plans is rough. The newer reasoning models with variable thinking budgets probably made per-request pricing untenable in the long run, but the way this lands on existing annual subscribers feels harsh. Going to look into the prorated refund.
I pay for github copilot solely for completions and use codex for actual agent work. I really wish I could pay like $5/month for just completions or there was a good local alternative for them.
What's the residual value of Copilot after the changes? For Enterprise plans, even copilot code reviews will be charged at token price + github action minutes for the execution of the review. One can roll one's own reviewer and have it spend the API tokens as well. At least one will be able to select the subset of files for the change set, if needed.
The background agents will also depreciate in value because of their harness that's a black box that's not optimized for token usage at all. Rolling one's own will be a better choice here.
Whose idea was this “premium request” model anyway? If you’re going to invent a new metric used to bill, why not align it with what, even at the time, was a clear underlying cost structure that GitHub actively chose to ignore for a more confusing system.
It made more sense in the ye old days where a request was basically just a chat message in a sidebar and it could also edit code. Then saying someone can use 300 chat messages a month kinda makes sense.
Turns out when a request can spawn tens of subagents and use millions of tokens over many turns of toolcalls then suddenly github copilot has a massive financial problem on their hands.
This approach started with the “Ask a question about your code” feature, which is more comparable to single chat message with relatively predictable token usage. Now it’s an agent who might work for 30 minutes, read the whole codebase, and write 1000 lines
I'm not usually a Conspiracy Guy, and the answer is probably `incompetence * tech_debt`. But I think that having sufficient layers of abstraction to any billing model is a useful way to hide the real cost of things. It's why it's done everywhere.
> Code completions and Next Edit suggestions remain included in all plans and do not consume AI Credits.
This is the VSCode autocomplete stuff right? Really enjoy this.
> Copilot code review will also consume GitHub Actions minutes, in addition to GitHub AI Credits. These minutes are billed at the same per-minute rates as other GitHub Actions workflows.
Looks like Microsoft has run out of compute and can't scale it fast enough to serve copilot users and Azure AI Foundry needs, given that the customer base is growing there as well.
I'm happy I invested in local solutions and cutting context to the bone for API providers. Claims about AI being able to fully replace programmers never took into account the long-run equilibrium price of inference.
That's my point. They made those decisions without any consideration for the long run, which would have required them to project the cost of AI services years into the future. Obviously management didn't do that and had no way to do that. It made current earnings look good, though, which was enough for them when they made the decision.
I only use copilot for the occasional auto-complete suggestion. I'm betting I could run a lightweight local LLM with llama.cpp to get similar functionality. Maybe this would be a decent replacement https://github.com/TabbyML/tabby
I was always curious how they can sustain a request based pricing model when requests can range from tiny to huge with all the modalities GH Copilot offers. Was a steal for agentic coding that turned out to be too good to be true, in the end. Still: Thank you for the ride.
I guess this is the "Google App engine" of Vibe coding when google raised pricing of App Engine significantly after being in preview. Doing weird price changes like this when coming out for new services make much more sense than doing this change which will just anger users.
According to `bunx ccusage` I'm easily doing $250-400/day in "real" API costs on my $200/month plan. There's no way everybody else isn't going to do the same thing and completely change the industry again. Both beginner and advanced developers are already hooked on all this stuff and they all know it.
Built credit pricing into my SaaS for AI features and the hardest part wasn't the math, it was that customers can't easily predict their own usage. They underuse and feel cheated, or overuse and churn. Subscriptions hide that volatility from the customer. Usage based pricing makes it their problem, which is honest but harder to sell.
We all knew this from the very beginning but couldn't compete with OpenAI or Anthropic on their subscription-based pricing strategy. It was nuts except for those few corporations burning investor money to keep competition out as long as possible. Now they don't have to hide anymore that subscription pricing won't do it for ai. The pyramid scheme is falling.
I'd be fine with it if they can make Sonnet 4.5 unlimited. I haven't personally seen any major differences between 4.6 or any of the other newer models. Claude 4.5 seems to be the right balance and works great for me.
The problem is that people expect to get the output of 100 people with a $20 subscription by spawning multiple agents. This is unrealistic. I'm using 2 codex plus account and able to manage a repo with 265-300k lines of code.
The point is - if its the same or more expensive per month than a real human employee - why pay for AI ?
Human retain knowledge, product knowledge, can pick up more work often for the same money. And having many of them means your business wont go down if provider suddenly bumps API pricing.
How much is the average pay for a junior developer in US? Its definitely much costlier considering PF and other benefits. Just the math. if you use it efficiently it's much cheaper than hiring a permanent staff. You can maintain a lean team and do all the mundane boilerplate coding with AI.
As a Github Copilot user, who mostly just uses chat in the VS Code editor but still burns through my Pro limit every month -- what's the best alternative price to performance? Claude Code?
"Paid for annual? Tough luck, from now on your usage limits are reduced by 89%. You can do 11% of what you paid for. Good luck if you paid annual a month ago!"
And then they have the gall to say
> "The bottom line: Plan prices aren’t changing"
If anyone lives in a place like Germany or Australia and has an annual sub, please take them to court, you're guaranteed to win because you have reasonable consumer protections and their ToS doesn't stand a chance. 9x reduction is unreasonable and the consumer cannot be expected to see this coming.
In case some diehard enshittifier believes that consumers should know better and businesses should be allowed to get away with it, where is the line? 99% reduction? Is that still okay?
If this situation is to be acceptable then it should be regulated as a financial product like stocks, which come with knowledge tests of "do you know you can lose all of your money?". And come with regulatory compliance and all that.
In a way, this is good news. I am expecting to see a lot less of AI-boosters claiming that they are "productive in ways you don't understand", now that they'll have to dish out upwards of a few thousand bucks per month for whatever PoCs they are building. And the excited hired middle managers and CEOs spending other people's money will have to be much more careful about their budget starting June.
It may already be cheaper if you calculate the hidden cost accumulated by LLM generated tech debt. At least some of it could be avoided by manually doing it better to begin with.
I searched for "Show HN" for the first time ever and I got the impression that most of the costs associated with the projects could have been avoided by never starting them to begin with.
"VibeBrowser", "Financial Database API for Vibe Coders" followed by dozens of self feeding AI projects with barely anyone doing anything that isn't related to AI.
so what's everybody using to get autocompletions in vscode? i've been using copilot just because $10 is cheap, but i use opencode for everything other than completions.
i tried the continue vscode extension, and it seemed kind of janky. are there better options?
I tried antigravity, it worked great for a couple weeks and then Google changed their rate limits and ruined it - I’d get about 3hrs of work before all the ai features just stopped working.
just cancelled mine. Only reason to have it was I could get a lot more done with a single prompt. No reason why I shouldn't go to the model provider directly instead.
This subsidized inference is just a marketing ploy to increase prices and profit.
If common people can have a DIY setup with an open source model cheaper than those behemoths with a scale advantage, it's clear that we have been played.
Time to either self host a Chinese open source model or to just pay the cheap Chinese providers.
Yeah, local is clearly the future. Even beyond the cheap Chinese models you can install the apfel[1] stuff if you're on a mac and want a quick available onboard cli option. And I'm sure people will adapt the Flash-MoE[2] integration to be even better soon as well.
DeepSeek (and other open weight models) can be served by anyone with the means to do it. All of the big open weight models are available on OpenRouter via commodity providers.
cursor, windsurf, and CC are all already on usage-based models so I guess what really matters is whether Copilot's GitHub integration depth justifies the price per token vs the alternatives
I'm not sure I understand this. All I know is now, I pay $39/month (actually less because I paid a year up front), use the agent, mostly on auto--and only choosing a model if it got stuck or in a loop--every day, and haven't hit any limits yet. It seemed to good to be true, after hearing others talk of $300/month bills. I guess it was.
Glad this was announced because I didn't even realize that our (small) team has been paying $20/month for Github Copilot when none of us are using it, so used the opportunity to cancel CoPilot altogether. I think it was free when I first activated it (this was also before Claude, Codex, Gemini, and while not that great, it was why not), and didn't realize it had switched to paid and bundled with our Github bill, and now per usage.
Re 1: Current models don't solve coding. They are useful tool for it though.
Re 2: Open weight models seem to be less than a year behind proprietary ones, so sure, if you're willing to spend tens or hundreds of thousands of dollars on a super computer that you probably don't fully utilize instead of renting time on someone else's super computer for a lot less.
People need to wake up and stop being surprised by these billing increases. I see it on every update of every model. This was all subsidized by VC and company money. Now they need a return and the prices will keep going up. Be glad that you took advantage of that up until now, but can we stop the pearl clutching when we all know the amount of money being dumped into AI and the lackluster returns?
It's less surprise, but more confusing given the game theory as their competitors are not doing the same thing and the multiplier changes alone will likely churn current users.
tldr: people were running multi-hour agentic coding sessions for the same flat fee as a one-liner autocomplete, github was eating the bill, and that party's over on june 1st