@OpenAI innovates stickers while car engine suck, burned 70% Pro 20x quota in 8h w Astra (XHigh), it was busy validate tests after handoff from Claude rather doing the planned work, while only 1% usage left on Opus 5.5 created beautiful concept prototype. Moon vs Sun experience.
Show Cursor Ultra usage with clear stats, run a cloud session for entire day with Opus 5.5, Sonnet 5.5 , Grok 4.7: - 9% left for other model, usage is definitely way above 400$ quoted anywhere. - 59% left for Grok is actually very disappointing. Not viable sub 🧐😤 @cursor_ai
@kunchenguid Interesting to see OpenAI and Anthropic playing on different games: the bottle and juice. Today OpenAI unleash nicer packaging of extensibility and accessibility but true judgement really on Sol 6.1 can really produce great result and efficiency. Ecosystem is new paradigm shift.
Since Claude released Opus 5.5, i used it for all projects and havent touched ChatGPT Pro sub, today I asked Astra (Max) to make a pitch deck, instead it made product review slides 😤. It’s no longer about usage gap but “intelligence gap”.
@thsottiaux Please address both quality, cost and performance truly as Anthropic did. Claude won me back so much, while ChatGPT:
- Desktop app is clunky, eat memory like nobody
- Astra even on Max, slow and dumper than Opus 5.5.
- Computer use is abused everywhere everytime.
@ashen_one It's very easy to get addicted to the feeling of usage maximization, we should focus on the quality output, if can build fully working app in a week and make real money from it. 10x usage no use if 2x the issues, time consumption to fix offset everything.
First impression, Opus 5.5 is smoking fast, straight to the point, if Astra can't slap anything meaningful later in review work then clearly big shift gonna happen.
@nick_white@cognition Im going to cancel Cursor Ultra and keep Devin Max next month, Grok is disappointing compare to SWE-2. Dollar per performance is no longer there for Cursor.
Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task.
In Claude Code:
- 5-hour session limits increase 20% today
- Opus 5.5 is priced lower, so it goes 25% further within limits
- Pro, Max, and Team users get a reset to use anytime
@OpenAI ChatGPT's usage and performance are really bad now. Pro 20x consumed 40% limit within a day without any coding, just code review and specs work as I move building with @cognition SWE-2, it also keep interrupted half through, first time I got it continuously.
@Davidwuuu92@Da7_Tech Indeed, Max 200$, Fusion (Fable/Swe-2). Running local got free promotion, but not the cloud agent. However, i got better result than Cursor Ultra which generous on Grok but not so much in other models (Fable, and no Astra)
@Davidwuuu92@Da7_Tech No, it does not appear to be that generous. I hit full quota on Max in 2 days running on cloud agent where they don’t offer free usage. So enjoy free promo on local build while it last.
Having fun with @bot and grilling with @cursor_ai while waiting for Claude and ChatGPT reset 🤠🤠.
Cool and fun stuff @poteto@roshan_s@mattyp
https://t.co/eCmOr4BRUU
https://t.co/x6GkKMjDzu