Calendaty

Opus 5.5 marked a huge step forward in AI, but its not enough We need smth that matches: - Opus 5.5 in intelligence, UI, taste, & thoroughness - 6.1 terra in speed & computer use - 6.1 sol in price The first lab to do this is gonna win & dominate

@jullerino holding up T3 Code & all the slop @theo merges

Codex limits on the $200 plan are a joke now In 1.5d I used 82% of my limit even tho I used Claude alot in the middle After the nerf $200 plans became super useless I hope other AI labs wont follow

Sure, Haiku 5.5 is nice But we need Fable 5.5, its nicer

GPT 6.1 sol high fast + Opus 5.5 high >> Opus 5.5 >> 6.1 sol

It finally happened After tons of work from the entire team v2 is finally merged Huge shoutout to @jullerino & the team for the amount of work he has put into this Should land on nightly in the next couple of hours ;)

With Fable 5.5 Anthropic better make it 100% of the weekly limit & make it more efficient

Compute is expensive, & thats bad for us: users OpenAI already nerfed $200 plans & introduced a $500 plan, other labs will follow Its bc compute is expensive, & they're running out of compute This means: - $20 plans: new free tier - $100 plans: new $20 - $200 plans: new $100 - $500 plans: new $200 This will get more expensive over time if compute doesnt get cheap

GPT 6.1 Sol is INCREDIBLE This model is very good, & its surprisingly limit friendly on $200 plans Running it on high fast across 5 threads with subagents only consumed 2% within 3h 🤯 The code quality is also pretty good Its nearly as good as Astra (even better) Overall this will replace Astra for me & make me use Opus 5.5 less (mainly for UI stuff)

You can sign in with your Codex sub in T3 Code directly Its official

This is a huge step forward towards mobile dev Imagine combining this with Opus 5.5's abilities in UI to make apps for the iPhone Duo I will share a guide soon on how you can go from idea → design → working app in a weekend The future is pretty promising

Opus 5.5 nails design After a few iterations on the AppStore assets, this is the result Its also really nice to use inside @paper, then convert to code I will share a vid soon on how to design & code REALLY nice landing pages & assets with Opus 5.5 & Paper

You can use Claude after hitting rate limits Anthropic secretly dropped this option alongside Opus 5.5 for some users In Claude Code, run "/low-priority" This will let you use Claude after hitting limits, but it'll be slower & only work if theres capacity at that time Enjoy!

Opus 5.5 is INSANE at UI Its the best I've seen that generates great UI I was working on some new AppStore screenshots for T3 Code, within an hour I made this (Still WIP) The best part is that its VERY good limits wise I LOVE this model, Anthropic is soo back

This is huger than the Opus 5.5 launch Banked resets in Claude, finally Thank you orange Tibo

Grok 4.7 is nearly as good if not better than Sol/Opus while being soo much cheaper This is crazy

ppl dont get what "frontier" is & how to correctly use it This post made me notice that most ppl doesnt understand whats frontier &/or not using it correctly You wouldnt: - feel the difference between Sol/Astra if your tasks arent hard - notice the limit difference if you're not shipping enough - push models to the limits if you dont ship real/complex stuff Thats why lots of ppl feel Astra is on par with Terra/DeepSeek, or that Astra isnt a big leap over Sol I will recommend watching the latest @theo vid for more But if this is you, go ship smth complex & ship more 🫡

Codex limits are ridicules In 5 days I finished: - 2x $200 plans - 1x $100 plan - 2x banked resets on the $100 plan Now I'm at 0% on all plans 2 banked resets used in 1 day & yet limits vanished $500/mo barely gets me 5 days of Codex This needs to be fixed

Tibo, its time for a reset I finished my limit on 3 Pro accs 😭 Limits on the $200 plans are sooo nerfed

Conspiracy Theory: OpenAI silently added peak hour limit I noticed that my Codex limit consumption is slow & good at some time While other times its VERY fast for some reason for the same workload, model, & everything Models also gets slower/faster We need confirmation