lalo

@pli_cachete Would love to hear your review regarding 5.6

@LololL5887 Better

Fable has triggered its filter on a geometric measure theory question 3 times in a row. I know you Ants can do better than this!

@pli_cachete I think Sol Ultra is the reasoning effort, and Pro is the different model ? ( harness ) ?

@pli_cachete LMAO AHAHAHHAHAH

I can finally talk about gpt 5.6 now. The poop and pee? Absolutely next level. And the blood and cum? I’ll be able to say more tomorrow…

@mertunsal2020 Amazing 🤩

Today, we are releasing Le Chaton L∃∀N, aka Leanstral 1.5. It achieves SOTA performance on graduate algebra benchmarks FATE-H and FATE-X and improves Pareto Frontier on PutnamBench, solving 587/672 problems with a x10 cheaper budget. 🧵

@patience_cave AHHAHAHAHAHAHAH LMA

i see how it is, when leo says it he gets a gold star and an ice cream, but when i say it everyone wants to kick me GPT-5.6 won’t drop next week either take that

@pli_cachete That’s good to hear

After I started feeling my emotions again, I realized that I was extremely lonely! I’ve since had regular calls with several very kind, wise people on here, and have 2 new people I am calling with this week. I highly recommend this! We are social creatures. Some of our best thinking/feeling comes from the space you occupy with another person.

@pli_cachete Your dog is so beautiful 🤩 what’s the name?

@pli_cachete 🥺🙏🏻

@teortaxesTex So excited

Yeah it seems we have DeepSeek-V4.1-Flash on web now. Announcement within hours-days I'd guess very different style (old vs new). And I think it's no good for web – when I gave it a console error, it Counted Every Line "manually". It is designed as a tool-calling agent first.

@pli_cachete I hope you are okay

@test_tm7873 Congratulations!!! Happy birthday!!

18th Birthday!!!! Yey!!!!

@CanopyWave_CW I had good experience with CanopyWave in general, and I'm really thankful for their team for always listening to the feedback, and being so responsive! 50M tokens is enough for the light user, but I'd recommend getting the 200M plan, for the moderate user.

I have been using the "Unlimited 50M" token plan by @CanopyWave_CW, powered by Hermes Agent. Kimi-K2.5 has been my main driver for this mini project and UI is quite good & the model itself has good vibes. It was essentially one-shotted, and even though there's lots of work to do

@pli_cachete Omg you finally tweeted

Some time in the last 3 months Sarah stopped being “Tripp’s girlfriend” and I became “Sarah’s boyfriend”

@CanopyWave_CW @CW_Gavin 🔥🔥 WOW!!!!! That’s awesome!

The first-ever #Unlimited Token Plan is now live. Real unlimited tokens, up to 90% lower cost than Claude. Unlock the full potential of your OpenClaw with Kimi K2.5, the best fit for complex multi-agent workflows. ✅Real Unlimited Tokens - Run all your agents without worrying about token limits ✅Privacy & Compliance Guaranteed - All inference runs are fully secure, protecting your data globally Stop paying high fees for Claude. Plans start at $15.99/month. No overage fees. No hand stops. Just seamless automation for your workflows. #OpenClaw #Kimi #UnlimitedToken #AIAgent

@CanopyWave_CW @CW_Gavin 🔥🔥🔥🔥

Built for daily coding workflows. The Canopy Wave Coding Plan gives developers access to multiple specialized coding models in one subscription: GLM-5 · Kimi K2.5 · MiniMax M2.5 · MiMo V2 Flash Use the best open model for each coding task. Starting at $15/month: · 500 requests/day · 10,000 requests/month Built for developers: · OpenAI-compatible API (works with Kilo Code, OpenClaw, OpenCode, etc.) · Optimized for large codebases and debugging · Stable inference with low latency · Secure AI infrastructure in Iceland and Canada Start coding with Canopy Wave today: canopywave.com/codingplan

@AcerFur Happy to hear honest reviews!!! The twitter hype makes me believe that these models are more capable than they are 😭😭

I’ve had access to GPT-5.4 Pro for a bit. What I will say on the vibes is that at backend code, it will happily write a few thousand lines, but can be somewhat buggy. On math, it has more creative ideas, but is also a lot less careful than GPT-5.2 Pro.

@AcerFur I hope GPT-5.3-Pro or whatever will be their experimental IMO Gold model

GPT-5.3 rumours on my timeline They know something I don’t clearly fr