Open source model usage is beginning to spike up! We are seeing a significant spike in - Qwen Coder - Kimi - GLM 4.6 usage While these models are behind closed source SOTA models. They are pretty useful for coding tasks

Haiku 4.5 - Comparable To Open Source Models, Good At Agentic Coding - solid agentic model, slightly better than GLM 4.6 - slow, as it uses a lot of test-time compute - isn't much more economical than Sonnet 4.5 - GPT-5 vibes Overall, it may be suitable for simple tasks, but we

Best Model Per Use-Case Presentations - Gemini 2.5 Full-stack apps - GPT-5 Codex, Sonnet 4.5 Docs - Gemini 2.5, GPT-5 thinking Videos - Sora 2 Images - Nano Banana Coding - Sonnet 4.5, Grok Code Fast Browser use - Sonnet 4.5 Doc Processing - Gemini Flash Enterprise Search -

Our launch tomorrow is about providing you with always-on, real-time cheat codes. After all, AI should be able to help you with every meeting, sales conversation, and interview in real-time 😎

So in the post AI world, AI will do everything BUT... you need to - buy A for creating your agent - buy B for the database - buy C for auth - buy D for hosting - buy E for connecting things - buy F for storing the memory of connected things ARE YOU KIDDING ME? THIS IS A

🚨 Claude Haiku 4.5 Dropped And Seems Fantastic!!! We will, of course, have it on ChatLLM, DeepAgent, and LiveBench today. But this may be the small model we have been looking for! Should unlock all our agentic workflows at scale 🚀🚀

Veo 3.1 Will Be In ChatLLM Later Today, But I Don't Get It Veo 3.1 Pricing - $0.40 per second Sora 2 Pricing - $0.10 per second Why should I be using Veo 3.1? 🤔

LLMs will literally do everything They will - vibe code your favorite startup - do your Xmas shopping - call your mom - post on your Twitter - answer your slack messages - and engage in erotica so you can save on OnlyFans! Yet, there are people who hate AI! 🤯

Meanwhile, In Enterprise Land - AI adoption is super slow - organizations are routinely blocked by "security" - "AI committees" stuck in analysis/paralysis - humans continue to do all the heavy lifting This is in STARK CONTRAST to the early adopter community 🤯🤯

We are living in a fantasy world. - non-stop demos of Gemini 3 (model not launched) - billions spent on AI slop (err.. creativity) - companies that are worth $30B with no revenue - millions of apps created by AI-assisted entrepreneurs THE PARTY WON'T STOP 😉💃

It’s not about GPT-5 OR Sonnet 4.5… It about using GPT-5 AND Sonnet 4.5 GPT-5 is an excellent planner, researcher and writer Sonnet 4.5 is an excellent do-er, coder and tester You need both of them to really leverage AI

Literally all in on AI! There will be no long term recession as there is still so much value to be realized

AI can transform mediocrity into brilliance People with ordinary skills can vibe with AI to be - an expert programmer - a top influencer - an artist - or good writer You can be a LITERAL GOD at what you do

People are ditching college and directly doing jobs or starting companies I am hearing stories of Stanford and Berkeley students starting companies as freshmen Some seniors are starting full time jobs even as the finish college They can’t wait to participate in the AI

Gemini 3 Fever Dream - it’s beats Sonnet 4.5 and GPT-5 Pro by a mile - it’s insanely fast - it’s stupid cheap - it everywhere al the time - it makes programmers obsolete It’s a bird, it’s a superhero, it’s an LLM The die hard fans won’t stop with the hype ❤️🔥

Literally all of open source AI is being propped up chinese developers