So after a very tepid response to Sol / Luna 6 - right when opus 5.5 is almost universally praised - OpenAI decided to cut in half their only subscription plan that was still decent value?
Things sure change quickly. 2 months ago you had nigh infinite usage and Sol 5.6 was a significant step forward. Now they're behind in both models and usage, and Astra 6.1 is too dangerous to release or some such (or possibly got embarrassed by opus). What are they even going to show off on today's devday, is a clone of grokbot really that interesting?
I prefer being able to get through a day without running out of usage.
I tried both sol and astra and they both decimated the $100/mo plan’s weekly limits in a single day running only 1 agent. Blew my mind. I’d always heard your subscription went further on OpenAI/Codex.
Comparatively I can run multiple opus agents in parallel almost continuously and stay under my limits.
I've always preferred GPT over Claude, but TBH the releases after GPT 5.6 Sol plus the surprise of cutting my usage in half is making me want to jump to anthropic and give Opus 5.5 a shot.
Looks like OpenAI is finally compute constrained (and they're already back to the market to raise money, presumably to for more compute).
I'm not sure if compute alone is enough of a moat enough but so far that's how they have survived Anthropic's capture of the enterprise/professional market. Along with shitposting on X and playing Anthropic's unforced errors on developer relations, safety downgrades and data retention.
But I suspect OpenAI will have to make some hard decisions about de-prioritizing their consumer products, which is most of their current valuation?
I wish they would give the notice one month in advance. Last week my $200 was renewed, and if tomorrow my usage cuts in half (if I understood correctly) it feels as I got partially ripped.
You're wrong for OpenAI, their 5x/20x always referred to the full weekly limit, and my (and others') multiple tests confirmed it empirically multiple times by calculating API-equivalent prices, especially when going from one subscription tier to another.
You're right for Anthropic though, their 5x/20x is mostly for 5h limit.
It's very obvious that they need compute for whatever they are announcing tomorrow and that thing will most likely have free usage. So they are cutting usage on the biggest users of compute to allow them to be able to serve whatever they are announcing tomorrow to the masses.
The thing is probably a Muse, and Grok Bot competitor.
Basically same move as Github Copilot a few months back when their golden rates ended.
They just refunded pro rata if you wanted out from the annual contract which they'd changed the terms on drastically, presumably OpenAI will offer the same to existing subscribers.
A shame that Tibo feels the need to BS like this. I predict another "we heard you...we handled comms badly...here's some resets" style post in a few days.
Codex $200 usage was already lower than $200 of Claude for around 2 months now. Half of that means that subscription is not really worth it unless you max things out all the time. But if you're going to go with API rates, there are better priced options out there than whatever OpenAI provides.
This is a risky strategy for OpenAI, and time will tell how will this work. But my hunch is that this is a catastrophic blunder.
I was listening to Dario's podcast where he was talking about how the entire game is in predicting extremely well how much compute you buy in advance, and if you miscalculate, it's pretty much over.
People are up in arms about this change but I think the knee-jerk reaction here is missing the bigger picture. API pricing is probably going down as a move to be more competitive in the enterprise space.
The ppl paying the actual bills have all been exploring crappier models to reduce costs and I think that has been strongly communicated to OpenAI.
Edit: People who think this is death for OpenAI need to look at Anthropic’s revenue. It has flatlined. Stingy decision makers are a bigger risk to these companies than stingy consumers.
What blow my mind is the fact that 20/30 dollar subscription is not enough to sustain the business? Pushing for 200 dollars subscription to keep afloat with it. If like claimed there are 1 billion users and 10% do the 30usd/month means 1 dollar/day x 100 million. How is that not enough to maintain this prices? Then you have all the inside business happening (users data, targeted ads, worldwide insights and intelligence, even monopoly of narrative) which would be your real passive income profit.
Users dont use all the time the service so they compensate each-other with power users. So for now getting from users at least 36B/year making an estimate evaluation of 360B (other incomes included, evaluation goes up to near a trillion). Be happy team you are the only one that you made it, but is enough one mistake in this house of cards. Again struggling to understand this logic even though you have access to the best intelligence in the world...
Because it's not 10%, it's ~25mn (tibo kept putting out resets at major milestones). It was like 4mn a few months ago.
The api margins are like 70%-90% (depending on who you ask, and not taking into account r&d), even lower with luna and their recent cuts. The subscription is even worse because they often use it as a loss leader for building hype at releases.
That's just not enough dollar volume to pay for all the inference, data center buildout, training, salary costs to compete with Meta etc etc. They also need to pay for the inference of the 98% of users who don't pay anything but use the models, that's essentially a marketing cost.
But Meta until very recently offered its services mostly for free, on a gigantic scale, paid for by ads. Now OpenAI has subscription revenue, API revenue, and ads revenue too. The fact that OpenAI isn’t financially sustainable makes me think it hasn’t really mastered its business yet.
They're literally out there hacking other companies and even governments and they're getting ZERO legal fallout from it, so I don't think legality is something they care about.
The beginning of the end of high subscription usage.
It was just matter of time for them to push everyone to APIs or pay more for the subs.
But the jump from 200 to 500 is much more than I expected.
They need to, because I'm leaving, going back to Anthropic for the time being. The last few days, their models feel dumb, thick.
They also no longer actively reset. The 20x subscription lasts a day. It is just a matter of time, before the consumer laws will catch up with these crooks.
I'm definitely getting more than 4 hours of Astra usage (medium or high, typically). However, I predominantly use Sol or Luna because Astra simply isn't (or shouldn't be) necessary for every task.
That said, I'm not able to get anything useful done in those hours. The exception is when I swap back to 5.6 Sol/Luna, which seem to be working as they used to.
Mostly, I've just given up and am using Opus 5.5 for all the important tasks now. Which is odd for me say, because I cancelled my Anthropic sub a few months back because I couldn't stand Opus 4.6 < x < 5.5. 5.5 feels great though, like Astra did when it first released.
I was really looking for the $500 plan as my $200 using Astra low goes away fast, but now it seems like I will just have a $500 plan that will do what the $200 plan does currently.
OpenAI were initially competing with Anthropic in $ value of tokens, and then number of tokens, and then the 5h limit removal meant you could buy multiple plans for huge continuous usage. This gravy train was never sustainable from a margin perspective and must have cost them billions.
It’s now going to be interesting how anthropic react - 5h limit and retain 15-20x $ value, or go the same was as OpenAI.
Anthropic have the revenue to subsidise longer. OpenAI is lagging behind Anthropic on the API revenue which is where Anthropic is making the majority of its money. I expect Anthropic will smell blood in the water and double down on their subsidies.
Yes very possibly. If you assume your competition will always exist then I think it makes sense to increase prices as they do even if you have the best business model.
So, they’re finally trying to end the subscription subsidies. Presumably ahead of Q4 so that their financials will look better going into next year for the IPO.
The assumption of the math is that you are not using Astra (which is hilarious). If you are not using Astra there is no point currently when Claude Opus 5.5 exists.
A lot of people say they'll switch to Anthropic, and that's fine, but make no mistake - both of those companies are losing money on their products. Both are looking to eventually extract much more from the market.
Now that "coding is solved" or whatever, it's time to pay the bill. I can easily imagine companies letting employees go, not because of greed, but to cover the AI spend.
In all fairness though, OpenAIs $200 plan was actually 20x compared to Plus, whereas Anthropic pulled this shady wording where it's only your session limit that is 20x while your weekly is only 10x.
But yeah, weird timing after a meh release and their competitor killing it right now.
I can believe that (20x is not 4x of the 5x plan) but I’ve been on the $100/mo Codex and the $200/mo Claude plans and I can run the whole week with multiple agents on Claude, mean while I burned through my whole week in 1 day on Sol. Yes, it’s half the price of my Claude plan but even if OpenAi actually gave 4x usage for the next plan up (which they just halved), it would still be less than the usage I get from Claude.
I used a reset and tried Astra and it also burned though my weekly usage in a day, with a _single_ agent.
I might be totally wrong but I wonder how much is due to context window. Codex felt like it was compacting every 5 seconds, that really screws your cache reads.
Yeah, Anthropic's marketing is shady. But x of y doesn't really give me any information. I have no clue what 1x is on a given day with either OpenAI or Anthropic.
In the end, the only thing that ultimately matters is how much total value are you getting each week. And Anthropic has been dominating OpenAI here for the past 2 months. Even with that 20x-but-really-2x thing.
> we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan.
So, so tired of living in this world full of people who lie and obfuscate at every possible turn. Just fucking say it - "we're cutting limits in half". Five words. Living slimeball.
Wow. The changes are one thing, but this is terrible terrible communication.
Tibo and other big OAI social media accounts have built a real cult of personality over the last few months and it seems to be getting to their heads. Comrade minister of propaganda announcing that rations will be cut is not a good look. There is still time to save face here, apologize, and just tell the truth plainly (we are capacity constrained and have to make compromises). Preferably before he goes from serious eceleb to corporate lolcow.
So they're halving usage for people who already paid $200 for the month, without a 30-day notice?
If so, that's totally a bait-and-switch.
So after a very tepid response to Sol / Luna 6 - right when opus 5.5 is almost universally praised - OpenAI decided to cut in half their only subscription plan that was still decent value?
Things sure change quickly. 2 months ago you had nigh infinite usage and Sol 5.6 was a significant step forward. Now they're behind in both models and usage, and Astra 6.1 is too dangerous to release or some such (or possibly got embarrassed by opus). What are they even going to show off on today's devday, is a clone of grokbot really that interesting?
I slightly prefer Astra to Opus 5.5.
It doesn't suffer the same verbosity issue and the writing is somewhat clearer.
The vision capabilities of Astra are a bit better I believe.
That said, Opus is 3x cheaper and the obvious winner when usage is considered.
I prefer being able to get through a day without running out of usage.
I tried both sol and astra and they both decimated the $100/mo plan’s weekly limits in a single day running only 1 agent. Blew my mind. I’d always heard your subscription went further on OpenAI/Codex.
Comparatively I can run multiple opus agents in parallel almost continuously and stay under my limits.
I've always preferred GPT over Claude, but TBH the releases after GPT 5.6 Sol plus the surprise of cutting my usage in half is making me want to jump to anthropic and give Opus 5.5 a shot.
The claudish is better. I no longer want to rip my own eyes out.
> we guarantee that over time you always get more work done and with an increasing level of quality
or your money back?
Its pure theft. Changing the rules.
>theft
Capitalism. It's spelled capitalism. I do not understand why people confuse capitalism with theft.
"it will net out at half the dollar in API spend compared to the old Pro $200 plan"
So effectively 50% less usage.
Looks like OpenAI is finally compute constrained (and they're already back to the market to raise money, presumably to for more compute).
I'm not sure if compute alone is enough of a moat enough but so far that's how they have survived Anthropic's capture of the enterprise/professional market. Along with shitposting on X and playing Anthropic's unforced errors on developer relations, safety downgrades and data retention.
But I suspect OpenAI will have to make some hard decisions about de-prioritizing their consumer products, which is most of their current valuation?
I wish they would give the notice one month in advance. Last week my $200 was renewed, and if tomorrow my usage cuts in half (if I understood correctly) it feels as I got partially ripped.
That seems like more than partially ripped.
In all that the one thing I don't understand is how they can just change their product for existing customers without even an email. Must be illegal.
They're vague on what they're selling in the first place. Even if certain subs are 5x or 20x, it's never really defined what 1x is.
You pay by the month, and they can change next month’s rules.
They're halving the $200 sub so it essentially becomes 10x from 20x
Having two basic Pro accounts - one from Anthropic - makes much more sense now.
20x was never about total usage, only 5h one. For 2x price you get 2x usage
You're wrong for OpenAI, their 5x/20x always referred to the full weekly limit, and my (and others') multiple tests confirmed it empirically multiple times by calculating API-equivalent prices, especially when going from one subscription tier to another.
You're right for Anthropic though, their 5x/20x is mostly for 5h limit.
> 20x was never about total usage, only 5h one
That was the case for Anthropic from what I remember, with everyone being pissed off about misleading marketing.
For OpenAI:
https://learn.chatgpt.com/docs/pricing
https://x.com/thsottiaux/status/2094254532020818191
> The Pro 20X is quite precisely 20X the usage of the Plus subscription, so it does exactly what it says on the tin.
Factually wrong. I mean how can 20x be about 5h usage when 5h limits were not part of the Pro plan for a long time.
It's very obvious that they need compute for whatever they are announcing tomorrow and that thing will most likely have free usage. So they are cutting usage on the biggest users of compute to allow them to be able to serve whatever they are announcing tomorrow to the masses.
The thing is probably a Muse, and Grok Bot competitor.
Same price, half the power.
Basically same move as Github Copilot a few months back when their golden rates ended.
They just refunded pro rata if you wanted out from the annual contract which they'd changed the terms on drastically, presumably OpenAI will offer the same to existing subscribers.
A shame that Tibo feels the need to BS like this. I predict another "we heard you...we handled comms badly...here's some resets" style post in a few days.
Codex $200 usage was already lower than $200 of Claude for around 2 months now. Half of that means that subscription is not really worth it unless you max things out all the time. But if you're going to go with API rates, there are better priced options out there than whatever OpenAI provides.
This is a risky strategy for OpenAI, and time will tell how will this work. But my hunch is that this is a catastrophic blunder.
I was listening to Dario's podcast where he was talking about how the entire game is in predicting extremely well how much compute you buy in advance, and if you miscalculate, it's pretty much over.
Looks like OpenAI miscalculated.
People are up in arms about this change but I think the knee-jerk reaction here is missing the bigger picture. API pricing is probably going down as a move to be more competitive in the enterprise space.
The ppl paying the actual bills have all been exploring crappier models to reduce costs and I think that has been strongly communicated to OpenAI.
Edit: People who think this is death for OpenAI need to look at Anthropic’s revenue. It has flatlined. Stingy decision makers are a bigger risk to these companies than stingy consumers.
Yes, Astra will have its API price halved while the $200 plan subscribers will have the same usage as before.
That's not enough.
There were rumors of a model called "Astra Minor".
I expect they will release this today.
If it performs slightly better than Astra and, considering the subscription cut, is 1/4th the price, it could be competitive with Opus 5.5.
Opus 5.5 appears to be around 1/3rd the price of Astra with a subscription.
Opus 5.5 runs circles around Astra.
You seem to be an employee of OpenAI, no?
Some things are just obvious.
Thanks for (not) confirming.
How are the investors happy with this though?
Will they even get their money back in 50 years at this point?
Genuine question, who has done the math on:
* how much investment they have gotten
* how much they need to make each year to give it back in X, Y years?
* how much they are currently making?
Is the end game one gigant megaAIcorp that you can't say no to? The chinese models seem to preclude that.
Have we collectively decided this a big game of investment-chicken, but of course we are smart enough to get out of it just before it crashes down?
"have all been exploring crappier models" - lol
What blow my mind is the fact that 20/30 dollar subscription is not enough to sustain the business? Pushing for 200 dollars subscription to keep afloat with it. If like claimed there are 1 billion users and 10% do the 30usd/month means 1 dollar/day x 100 million. How is that not enough to maintain this prices? Then you have all the inside business happening (users data, targeted ads, worldwide insights and intelligence, even monopoly of narrative) which would be your real passive income profit.
Users dont use all the time the service so they compensate each-other with power users. So for now getting from users at least 36B/year making an estimate evaluation of 360B (other incomes included, evaluation goes up to near a trillion). Be happy team you are the only one that you made it, but is enough one mistake in this house of cards. Again struggling to understand this logic even though you have access to the best intelligence in the world...
Because it's not 10%, it's ~25mn (tibo kept putting out resets at major milestones). It was like 4mn a few months ago.
The api margins are like 70%-90% (depending on who you ask, and not taking into account r&d), even lower with luna and their recent cuts. The subscription is even worse because they often use it as a loss leader for building hype at releases.
That's just not enough dollar volume to pay for all the inference, data center buildout, training, salary costs to compete with Meta etc etc. They also need to pay for the inference of the 98% of users who don't pay anything but use the models, that's essentially a marketing cost.
But Meta until very recently offered its services mostly for free, on a gigantic scale, paid for by ads. Now OpenAI has subscription revenue, API revenue, and ads revenue too. The fact that OpenAI isn’t financially sustainable makes me think it hasn’t really mastered its business yet.
This doesn't bode well, 500$ is the new 200$? Wild that they're doing this after Sonnet/Opus 5.5 releases.
I wonder how legal it is to change usage during a running subscription.
> I wonder how legal it is to change usage during a running subscription.
Legality is no concern for these magical companies. Money is of no object, laws do not apply, the return on investment must not be spoken.
They also halved API prices with GPT-6, so I think it’ll net out pretty close for sol and Luna, but much less to be used for Astra.
They're literally out there hacking other companies and even governments and they're getting ZERO legal fallout from it, so I don't think legality is something they care about.
OpenAI only does monthly payments, so they can just keep the old usage for current subscribers for through their billing period.
The beginning of the end of high subscription usage. It was just matter of time for them to push everyone to APIs or pay more for the subs. But the jump from 200 to 500 is much more than I expected.
They need to, because I'm leaving, going back to Anthropic for the time being. The last few days, their models feel dumb, thick.
They also no longer actively reset. The 20x subscription lasts a day. It is just a matter of time, before the consumer laws will catch up with these crooks.
Well even if they open it - it'll still be that way.
With Astra 6.1 being delayed [1] then it'll be interesting to see what they offer today, otherwise it's Opus 5.5 until they get their stuff together
[1] https://gizmodo.com/openai-cancels-release-of-gpt-6-1-astra-...
If 20x lasts a day for you, that’s 6k a month, why not buy your GPUs and train them, deploy them?
Burner account? Nice. Another OpenAI troll.
Edit: You clearly need to go back to school, learn basic arithmetic.
> The last few days, their models feel dumb, thick.
My local Qwen 3.8 27B seems faster and more intelligent than new Luna.
Codex GPT-6 models are so congested and slow at the moment it is difficult to get through the current 20x plan anyway.
What do you mean? I blow through my 20x in 4 hours of Astra usage a week.
I'm definitely getting more than 4 hours of Astra usage (medium or high, typically). However, I predominantly use Sol or Luna because Astra simply isn't (or shouldn't be) necessary for every task.
That said, I'm not able to get anything useful done in those hours. The exception is when I swap back to 5.6 Sol/Luna, which seem to be working as they used to.
Mostly, I've just given up and am using Opus 5.5 for all the important tasks now. Which is odd for me say, because I cancelled my Anthropic sub a few months back because I couldn't stand Opus 4.6 < x < 5.5. 5.5 feels great though, like Astra did when it first released.
I was really looking for the $500 plan as my $200 using Astra low goes away fast, but now it seems like I will just have a $500 plan that will do what the $200 plan does currently.
Title should be - 'OpenAI cuts Pro $200 usage by 50%'
OpenAI were initially competing with Anthropic in $ value of tokens, and then number of tokens, and then the 5h limit removal meant you could buy multiple plans for huge continuous usage. This gravy train was never sustainable from a margin perspective and must have cost them billions.
It’s now going to be interesting how anthropic react - 5h limit and retain 15-20x $ value, or go the same was as OpenAI.
Anthropic have the revenue to subsidise longer. OpenAI is lagging behind Anthropic on the API revenue which is where Anthropic is making the majority of its money. I expect Anthropic will smell blood in the water and double down on their subsidies.
Yes very possibly. If you assume your competition will always exist then I think it makes sense to increase prices as they do even if you have the best business model.
> Over time, we see prices go low enough that it makes sense for most to buy usage as needed
Hard to believe when all they've done so far is raise prices and reduce usage quotas.
The 20x plan right now lasts about 1 day and they're reducing it to half a day of usage?
Put differently, they're trying to kill subscriptions. The old bait and switch.
I agree with them. It will happen circa 2035, when OpenAI will at best be a Microsoft division, acquired for $20bn.
And renamed to Copilot
Yeah... on OpenRouter or equivalent
For most work now it's already cheaper to buy tokens.. of DeepSeek v4.1 Flash from the official DeepSeek provider.
Yeah $20 there lasts forever
So, they’re finally trying to end the subscription subsidies. Presumably ahead of Q4 so that their financials will look better going into next year for the IPO.
Going to be hard if you've lost all your customers due to your questionable behaviour.
The assumption of the math is that you are not using Astra (which is hilarious). If you are not using Astra there is no point currently when Claude Opus 5.5 exists.
Not going back to anthropic as i love the alignment of 5.6 sol, but I'll start looking around
A lot of people say they'll switch to Anthropic, and that's fine, but make no mistake - both of those companies are losing money on their products. Both are looking to eventually extract much more from the market.
Now that "coding is solved" or whatever, it's time to pay the bill. I can easily imagine companies letting employees go, not because of greed, but to cover the AI spend.
Well, this is upsetting. I was already maxing out two $200 subscriptions days before the weekly limit. I wonder how bad it'll be after this change.
In all fairness though, OpenAIs $200 plan was actually 20x compared to Plus, whereas Anthropic pulled this shady wording where it's only your session limit that is 20x while your weekly is only 10x.
But yeah, weird timing after a meh release and their competitor killing it right now.
I can believe that (20x is not 4x of the 5x plan) but I’ve been on the $100/mo Codex and the $200/mo Claude plans and I can run the whole week with multiple agents on Claude, mean while I burned through my whole week in 1 day on Sol. Yes, it’s half the price of my Claude plan but even if OpenAi actually gave 4x usage for the next plan up (which they just halved), it would still be less than the usage I get from Claude.
I used a reset and tried Astra and it also burned though my weekly usage in a day, with a _single_ agent.
I might be totally wrong but I wonder how much is due to context window. Codex felt like it was compacting every 5 seconds, that really screws your cache reads.
Yeah, Anthropic's marketing is shady. But x of y doesn't really give me any information. I have no clue what 1x is on a given day with either OpenAI or Anthropic.
In the end, the only thing that ultimately matters is how much total value are you getting each week. And Anthropic has been dominating OpenAI here for the past 2 months. Even with that 20x-but-really-2x thing.
[dead]
Do note, it does not say 20x.
> we are also changing how we calculate the usage for it. In effect, if you do the math, it will net out at half the dollar in API spend compared to the old Pro $200 plan.
So, so tired of living in this world full of people who lie and obfuscate at every possible turn. Just fucking say it - "we're cutting limits in half". Five words. Living slimeball.
Wow. The changes are one thing, but this is terrible terrible communication.
Tibo and other big OAI social media accounts have built a real cult of personality over the last few months and it seems to be getting to their heads. Comrade minister of propaganda announcing that rations will be cut is not a good look. There is still time to save face here, apologize, and just tell the truth plainly (we are capacity constrained and have to make compromises). Preferably before he goes from serious eceleb to corporate lolcow.