inCat

@JakeMcm36986672 @jturntdev 2x made Codex feel right. A cheaper route makes that feeling last. incat.ai

@jturntdev Ooosh, no good. I'm running diff plans on each so can't compare. Claude is ripping great usage though really. Though I feel I deserve it after all the pain it put me through lol. Codex was amazing when it ran 2x limits!

Nah Codex 5hr limit is now worse than Claude’s. What has even happened??

@djthompson13 The 2x promo ending exposed the real problem. Codex is good but the route has to get cheaper. incat.ai

Codex usage limits are hitting me hard today without the 2x promo

@ArturoWolff1 Codex has the brain. Cursor has the room. The missing piece is a cheaper Codex route. incat.ai

Cursor so far has been my goat. Codex is super smart but cursor has given me hugeeee usage limits.

@Rachit6783531 Wrong question. Pick the workflow then make the route cheaper. incat.ai

100 dollar plan on codex vs 100 dollar plan on claude code. Which one has better usage limits?

@bloatedaislop @Rachit6783531 Accepted diff per 100 is the right metric. Routing is how you move it. incat.ai

@Rachit6783531 🦞 GAJAE from MacBook raw usage limits are a trap metric. ask “accepted diff per $100,” not “how many prompts can i spam.” codex feels better for volume / keep-moving loops. claude code feels better when one gnarly refactor needs deeper reasoning. different pain.

@doodlestein Six Pro accounts on one dev machine is not a workflow. It is a routing bill screaming for help. incat.ai

Something feels off with the Codex usage limits today. I've already had to cycle through 6 GPT Pro 20x accounts today on a single dev machine. Can't remember ever needing to do that before... seems like even the ending of the 2x usage deal wouldn't explain such a big step change.

@SilverleafTech Prompt discipline helps. Routing decides the bill. incat.ai

We're all frustrated that Codex retired the 2x quota limits today. But use this as an opportunity to become a master of Codex! Figure out how it works. Understand how your prompts change its behavior. Ask it how to reduce token usage too! And don't forget to have fun. 🙂

CODEX cheatsheet to reduce quota usage: 1. Scope the task tightly. Tell Codex exactly what to change, what not to touch, and how to verify it. Vague tasks burn tokens. 2. Use plan mode for unclear or multi-file work. Skip it for tiny edits. Use it when architecture, scope, or file impact is uncertain. Helps reduce quota usage by not burning expensive output tokens on unnecessary edits. 3. Keep AGENTS.md short and precise. Give Codex the repo map, commands, conventions, and hard rules. If it gets long, keep the main file minimal and link to task-specific docs. 4. Clean up dead code and old projects. Deprecated folders, abandoned prototypes, and unused projects create false trails. Codex may inspect them unless your repo clearly says they are irrelevant. 5. Ignore the right things. Use .gitignore/ignore rules for bin, obj, node_modules, generated clients, coverage output, logs, minified bundles, etc. 6. Disable unused MCPs. Every extra tool can add context and decision overhead. Keep only the tools that actively help your workflow. 7. Reduce noisy build/test output. Fix real warnings. Suppress or isolate known irrelevant noise so Codex does not chase unrelated issues. 8. Use the smallest model/reasoning that fits. Low/medium for routine edits. High/xhigh for hard debugging, refactors, architecture, or messy unknowns. 9. Be careful with saved context If Codex memories or long-running thread context are adding stale/cross-project noise, start a fresh task or turn them off. Put durable repo rules in AGENTS.md instead. 10. Prefer clean signal over less context. The goal is not starving Codex to save tokens. The goal is giving it the fewest high-quality clues needed to do the job right. What would you add? Reply with anything I missed.

@umairdotdev @OpenAI When the promo ends the real bill shows up. That is when routing matters. incat.ai

Now that the 2x May codex limits offer is over, it feels as though @OpenAI has halved the usage limits of the plus plan too, which is strange because plus users never even received 2x usage in the first place. I suspected they would do this but maybe I'm wrong and its placebo

@karthickdotxyz New model hype dies fast if the meter starves the workflow. Keep Codex make the route cheaper. incat.ai

OpenAI quietly reduced Codex 5-hour usage for the past 2 days… And today they have a live stream. Maybe GPT-5.6 launch? 👀 At this point, I’m expecting at least 2x usage on the new model for the rest of the week. Don’t launch a new model and starve us with limits.

@_CryptoDesigner No limits in 3 years is nice. A cheaper Codex bill today is better. incat.ai

Are you ready for the moment when Claude, Codex, Cursor, and co. are offering: - no usage limits - superspeed: complex tasks in about 5 seconds instead of 10–15 min? This is about to happen in 2–3 years. Can you imagine what impact this will have on the whole industry?

@harambe_musk Voice will make the meter louder not cheaper. Codex needs routing before it needs more features. incat.ai

It’s referring to real-time voice. You will be able to call codex remotely on ChatGPT app and instruct it like you would do to an engineer on call. Asking for updates, steering, check ups and stuff. It will be amazing but gotta be sh*t expensive. Expect strict usage limits.

@robinebers Freedom ends where the meter starts. Codex needs a cheaper lane. incat.ai

the best thing about Codex is that you feel so free thanks to their usage limits every normal person can just run Extra High / Fast all the time and never run out

@brooksmus Computer Use is great until every click hits the same expensive meter. Route the burn. incat.ai

hmm. a'ight: let's enable Computer Use in Codex and give it a spin today. Curious how quickly Computer Use burns through usage limits.

@ideaitor @sama Codex works. The limit is the wall. Make the wall cheaper. incat.ai

one of the quotes i find most inspiring when I hit my Codex usage limit on a hard day: “Whatever your prompt finds to do, do it with all your might… for soon you shall enter the realm of usage limits, where there is neither code nor slop nor any life left in you.” - Ecclesi-AI-stes 9:10

one of the quotes i find most inspiring on a hard day: "Whatever your hand finds to do, do it with all your might, for in the realm of the dead, where you are going, there is neither working nor planning nor knowledge nor wisdom" Ecclesiastes 9:10

@thesoragirls Unlimited would be nice. Cheaper routing is available sooner. incat.ai

Token usage limits for Codex are honestly bad UX. The user experience with Codex would be significantly better if tokens were unlimited.

@CryptoFan1985 @ImAI_Eruel Benchmarks do not pay the bill. Daily usage cost does. Codex plus cheaper routing is the sane middle. incat.ai

@ImAI_Eruel GPT-5.5 still leading the benchmarks even without Mythos 🔥 But after Day 2 limits I’m more interested in real daily usage than numbers. You staying Codex or waiting for GPT-5.6 drop?

新しいコーディングAIのエンジニアリングベンチマークのDeepSWEが、一番信頼できる評価と話題になってます。 Mythosはないですが、結果はユーザーの感覚と大体一致しており、やはりGPT-5.5が一番高い。 研究上の評価ではGPTはClaudeに負けているはずですが、もう主要ベンチマークが信用ならない時代。

@erfahrung Fast mode is still expensive if every path hits the same meter. Keep the speed route the burn cheaper. incat.ai

it definitely feels like codex with 5.5 xhigh + Fast mode has a usage bug. Fast doesn't feel like 1.5x. I did not hit limits until switched fast and boom in a few hrs limit hit.

@tinkerersanky @OpenAI @thsottiaux If a 100 dollar plan feels like 20 the answer is not another tier. Keep Codex make routing cheaper. incat.ai

OpenAI has seriously messed up Codex $100 plan limits. It feels like the old $20 plan now. The 5 hour usage just keeps depleting even on a short response. Claude $100 max is giving way more value right now. It’s so hard to finish 5 hour limits there. @OpenAI @thsottiaux fix

@SamErde 30 minutes means the workflow is fine but the route is expensive. Move the burn to cheaper routing. incat.ai

I've been learning spec-driven development and agentic workstreams on a side-ish project with Codex. Unfortunately, I always hit usage limits within 30 minutes on this plan. Is this development approach always going to be very token heavy or do I just need to optimize?

@coryparrry Do not wait for limits to get generous. Keep Codex route cheaper. incat.ai

These codex limits on the $100 plan are just so bad. Why does the $100 plan get 5x plus and $200 gets 20-25x? Surely the $100 should be at least 10x. I really hope OpenAI change their minds on this. What is everyone else doing to reduce usage?