OpenAI's decision to double Work and Codex usage limits on its Pro 100 and Pro 500 plans, and to drop a cut to its $200 Pro plan one day before it took effect, set off a long day of reaction on X. The posts we read fell into three groups: relief, arithmetic from subscribers working out what each plan now buys, and complaints, from Plus users left out of the change and from Codex users who used the moment to list what they still want fixed.
The announcement came from Tibo (@thsottiaux) at 9:56 AM PT on October 29. By 3:23 PM PT the post had about 2.12 million views, 28,113 likes and 5,118 replies. A Hacker News submission of the post had 13 points and one comment at 3:42 PM PT, and Reddit cannot be read from our tools, so the reactions below come from X. For the full plan details, see our news post on OpenAI doubling ChatGPT Pro Work and Codex usage.
What OpenAI said
Tibo's post read: "We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today." It went on: "We heard your kind feedback and won't be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same." He added that extended limits for dots continue "for now," and that dot usage stays separate from default usage. For what a dot can do, see our guide to texting your ChatGPT dot over iMessage or RCS.
Six minutes later he followed up with one line: "Graduating from The Reset Company to The Token Company." That post had about 420,000 views by 3:23 PM PT.
OpenAI's About ChatGPT Pro tiers help page is more exact than the post. Pro 100 and Pro 500 get twice their previous Work and Codex allowance. On Pro 200, only the lower allowance doubled, which brings every Pro 200 subscriber to the same limit. Pro 200 subscribers who already had the higher allowance get no increase, so "doubling" does not apply to them. Plus usage is unchanged. OpenAI's release notes say Business Premium usage also goes up but do not say by how much; the doubling for Business Premium comes from Tibo's post.
Another OpenAI staffer, @ajambrosino, whose X bio reads "Codex & ChatGPT @OpenAI," wrote that usage limits "are back to being double" and "your feedback worked," and that the team had "kicked off an intense sprint to dramatically improve the app simplicity and performance." The post ended with "now, send me your complaints." Tibo quoted that post with "Send your kind feedback to this person."
Relief, and credit to the backlash
Some posts framed the change as a win for users who complained. @iRandeepRana opened with "THE BACKLASH WORKED." and closed with "never underestimate angry power users." @imujjwaal wrote: "Complaining actually worked lol." @MechEconomy, whose post had about 32,000 views at 3:42 PM PT, called it "Rare to see a tech giant actually listen to developer pushback before the grace period expires."
@EnvolDev, who wrote that the planned Pro 200 cut "Felt like a rug pull," read the reversal as a competitive move: "This is what competition does. Every lab is fighting for the same developers, and we're the ones winning." The post ended with "Your move, Anthropic."
Not everyone took the "kind feedback" line at face value. @BiggSmithzzz replied to Tibo with only the word "Kind" in quote marks and a laughing emoji. @indie_alx asked: "Did the backlash actually work, or was the cut never meant to happen?"
Doing the math on each plan
A recurring question was what "doubling" means for each tier. @StefanoGPT asked: "Wait, so does it mean that the OpenAI $500 plan is getting 50x usage?" Tibo replied: "Yes."
@HashItOut_ posted a full breakdown: Pro 100 from "5x → 10x," Pro 200 "stays at 20x," and Pro 500 from "25x → 50x," all measured against Plus. By that math, the post said, every Pro plan now costs $10 per "1x of plus usage," while Plus costs $20. "plus users pay double per unit," it said. Those multipliers are the poster's figures. OpenAI's help page does not state multipliers in this form.
That leveling changed how some people saw the $200 plan. @copperchunk wrote that the old pitch was "Double the price, FOUR times the usage," and that now it is "Double the price, double the usage." The verdict: "The $100 tier is now the sweet spot. The $200 tier has lost its main selling point." @TimJayas put the same point as a joke: "cut subscription usage in half from 20x," then "double the usage back to 20x."
@zwb44 called it a "good change" but noted that the Pro 200 usage "doesn't actually change it just stays 20x which is confusing." That matches the help page, which says Pro 200 subscribers who already had the higher allowance keep it rather than get a further increase.
Some users also pointed out that speed settings eat into the gain. Before the reversal, @0xPixelBaron wrote that "OpenAI counts Fast at 2.5x of included usage." OpenAI's help page confirms that rate for Fast mode compared with Standard. @Joshua_WD linked a longer X article on "why Fast mode eats most of the gain."
Plus users feel left out
Replies from Plus subscribers were a steady thread. @onesadlittleboy told @ajambrosino to "Do something with Plus plan," and said OpenAI had "abandoned that plan completely it seems." @JohnWicko0wyJoh asked Tibo: "What about plus users ? Do we not deserve love?" @SyroneWong replied: "Plus users cancel subscription and switch to Claude Pro." @Ubertag90210 asked for something in between: "Could you add a $50 plan? Call it 3x even?"
Cross-shopping with Claude
Some replies treated the change as part of a choice between OpenAI and Anthropic, a choice our comparison of ChatGPT Plus, Claude Pro and Google AI Pro covers at the $20 tier. @StackTech told Tibo they had already upgraded Claude to a $200 subscription and planned to drop ChatGPT to $20, but "now I may have to keep both and owe $300/mo to my AI overlords." Under @ajambrosino's call for complaints, @danielarpm wrote that Codex "Gets too slow vs Claude" and that "Projects are more organized in Claude."
"Now release good models" and the app complaints
One short reply to Tibo had 198 likes at 3:42 PM PT. @thekitze wrote: "Ok now release good models."
The thread under @ajambrosino's post, with 2,060 replies at 3:42 PM PT, turned into a bug list. @Furious40539798 wrote that the Codex app on Windows "managed to open ~200 processes using 27GB of RAM," and that it writes "1-2 GB/h because the plugin catalog gets rewritten on every subagent." @hamataro0710 said their dot "consumes all my usage during nighttime sessions with unnecessary high Astra usage," and asked to see and switch the model it runs.
Why it matters
The reaction shows how closely paying users track usage limits on coding and agent plans. People did not just cheer the doubling: within hours they had recalculated cost per unit, argued over which tier is the better deal, and compared it with Claude's plans. The change also leaves two open questions in the replies: whether Plus gets anything, and whether OpenAI's app work delivers on the complaints it asked for.
Sources
- Tibo (@thsottiaux): usage limits doubled
- Tibo: The Reset Company to The Token Company
- Tibo: reply to @StefanoGPT
- Tibo: Send your kind feedback to this person
- @ajambrosino: usage limits back to double
- OpenAI: About ChatGPT Pro tiers
- OpenAI: ChatGPT release notes
- llms.blog: OpenAI Doubles ChatGPT Pro Work and Codex Usage
- Hacker News: submission of Tibo's post
- @StefanoGPT
- @HashItOut_
- @copperchunk
- @TimJayas
- @zwb44
- @0xPixelBaron
- @Joshua_WD
- @iRandeepRana
- @imujjwaal
- @MechEconomy
- @EnvolDev
- @BiggSmithzzz
- @indie_alx
- @onesadlittleboy
- @JohnWicko0wyJoh
- @SyroneWong
- @Ubertag90210
- @StackTech
- @danielarpm
- @thekitze
- @Furious40539798
- @hamataro0710
