i hit rate limits so you don't have to. first to post when claude, codex or cursor change usage limits or prices. daily copy-paste fixes.
Waiting for my limit to reset
Joined June 2026
- Tweets483
- Following80
- Followers40
- Likes166
Pinned Tweet
claude code, codex and cursor keep quietly changing usage limits and prices.
i post every change here within minutes, checked against the official source, plus one fix a day so you hit the wall less.
follow if you've ever been rate limited mid-thought
1
2
116
fable 5.2 cost $8.52 for the same steam train opus 5.5 drew for $3.80. paying 2.2x to get fog
newest model by default or cheapest one that works: a or b
Claude Fable 5.2 vs Claude Opus 5.5
Same prompt. Steam train on a snowy viaduct at dusk.
Fable 5.2: $8.52 · 20m 40s
Opus 5.5: $3.80 · 20m 01s
Opus got it in one round. Fable's first try came out murky, the fog hid the mountains, so I sent it back once.
After the fix it caught up. Fable has the heavier blizzard and the frozen river. Opus has the two-tier viaduct and real mountains behind it.
2.2x the price to end up even.
What do you guys think?
12
11% of a claude code session limit for clicking back into an 806k token chat after 4 hours. the cache went cold, the meter did not
/clear and lose the context or eat the reread: a or b
came back to a claude code session after 4h 13m idle.
it was 806k tokens long so it had to read the whole thing again from scratch.
that one click was 11% of my session limit :)
2
2
81
$4k of opus 5.5 to build something named mailcheap. the only cheap part was the name
would you eat a $4k agent bill to ship a product: yes or hard no
It cost me ~$4,000 to build MailCheap with Opus 5.5 on @cursor_ai so ironically, it's my first unprofitable product.
17
gemini 4 hydrogen is out. 2.5x faster than 3.8 flash, live on every ai studio tier, but antigravity gets it on paid pro and ultra only. fast new flash or the model you already trust with your repo?
Meet Gemini 4 Hydrogen: built for rapid coding, multimodal synthesis, and everyday workflows.
It finishes tasks 2.5x faster than 3.8 Flash and surpasses Argon on the @ArtificialAnlys Intelligence Index.
24
claude code getting slower and pricier the longer you ask it stuff?
every quick "wait what does this file do" stays in the conversation for the rest of the session, and token costs scale with context size
ask it on the side instead. /btw answers about the current session without adding to the conversation
what eats your context first, side questions or giant file reads?
1
1
25
google just shipped a model that makes up to 1000 decisions per call (gemini decision 1, in ai studio and the api)
i make one a day: which tool i burn my limit on first
would you let it pick your model or your merges
introducing the Gemini Decisions API
powered by Gemini Decision 1, our new multimodal decision model
get fast, calibrated decisions on up to 1000 questions at once across text, images, audio, video, and PDFs
try it today in AI Studio and via the Gemini API ai.studio
26
codex ignoring the AGENTS.md rules closest to your code?
it stops loading instruction files at 32 KiB combined, root first. so the folder-level ones get dropped
fix: trim, split across subfolders, or raise it in ~/.codex/config.toml
project_doc_max_bytes = 65536
trimming yours or raising the cap?
1
16
claude just gave pro, max, ultra and team one free full limit reset, usable anytime before dec 10 at 2pm pt. burning it tonight or saving it for a bad week?
46
max and ultra people, day two of fable 5.2: fable for everything, fable to plan and opus to build, or still all opus?
17
chatgpt business seats now sign into opencode, amp, kilo code, warp and a bunch of other tools.
does this replace your api key bill or not?
Now, you can use your ChatGPT Business plan in tools like @NotionHQ, @v0, @ampcode, @kilocode, @warpdotdev, Hermes from @NousResearch, @openclaw, @opencode, and more.
Connect with Sign in with ChatGPT to get started.
26
claude code eating your limit after a break?
your first message after the prompt cache expires reprocesses the whole conversation
run /usage and read the prompt cache line. it shows misses, the likely cause, and if the cache is still warm
switching to unrelated work? /clear first
21
gemini 4 carbon is ultra subscribers only in the gemini app and antigravity. api and ai studio have it today.
paying for ultra to try it, or waiting for pro?
Starting today, our next frontier model Gemini 4 Carbon is rolling out to Google AI Ultra subscribers on desktop, mobile, and web at gemini.google
It’s our strongest model, built to reason deeply through your most complex work:
✅ Get more thorough, reliable outputs on your toughest tasks
✅ Turn dense data and docs into polished presentations, spreadsheets, and ready-to-use reports with more precise analysis
✅ Create web apps, prototypes, and 3D scenes with stronger spatial understanding for 3D and UI design
1
46
claude fable 5.2: max/ultra get 1-3 free fable limit resets based on how long you've paid (under 6mo 1, 6-12mo 2, 12mo+ 3). pro gets fable for a week, up to 50% of weekly limit, plus 1 reset. it's in settings > usage
burn it today or save it?
1
110
claude fable 5.2 is live in claude code and the api. $8/$40 per 1M tokens, down from $10/$50, and up to 30% faster
20% cheaper per token. does that actually stretch your plan or just your api bill?
1
1
46
“Review everything I’ve done and make it better” make no mistakes.”
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
6
claude code eating your limit faster than expected?
1. run /usage. on a pro or max plan it shows what your usage went to: skills, subagents, plugins, MCP servers
2. read the recommendations under it
3. press f to let it apply the ones it can
the f shortcut needs a recent version
1
60
claude code hit a usage limit mid-task?
type /rate-limit-options before you bail
wait and continue when it resets, add usage credits, or upgrade. on a claude.ai plan it can open this menu itself
2
2
107