Welcome back, OpenAI ! Day-0 support llama.cpp with MXFP4, let it rock 🚀🤘

😱Unreasonable efficiency of GPT-OSS-20B reasoning trace 😱 Yeah… it’s really good. This is exactly what we wanted to achieve with the Lucy model, a natural and effective reasoning trace that is less prone to hallucination. Well, Lucy is a 1.7 B model after all, so it’s

Alan Dao Ivan Fioravanti ᯅ It’s too bad that M-chips can't use FP4. To achieve the same quality of output as a Blackwell chip, Apple requires four times the memory. This is yet another strike from NVIDIA's software ecosystem monopoly.

Shocking insight - GPT5 has a 25% improvement over GPT4

Claude Caude can run GPT5. GPT-5 is good at fixing nasty bugs and doing research. In a different way than Claude and they can work together. Just create a Claude code agent called gpt5:

Unlike an RNN, one attention block alone cannot model anything interesting. And it’s the stacking of it that does wonders. Understanding this compositionality should be at least as important as understanding the attn module itself.

Haha lmao cool response 😂 But deep research is overkill for simpleQA don't ya think??

👋 Jan Fkn unreal, 91% 🤩🔥? You guys keep delivering! Is this you as well, Alan Dao ?

Pretty cool. I think 2025-2026 will be a stronger focus on these in open source tooling. I.e. having LLMs delegate knowledge-based queries to search, which in turn frees up model capacity to improve reasoning capabilities and tool use.

I really really like 👋 Jan It's a very friendly app to locally run LLMs, great for privacy I've tried others like LM Studio and Ollama and they're nice but very engineer-built, a bit too difficult for me Jan is simple and cute and pretty and a great alternative to talk to

How about 100% SimpleQA 🤣

Jan v1 is trending on Hugging Face today. Huge thanks to everyone trying it out, giving feedback, and sharing your setups. We see you 💙

This is a good idea. We have similar observations in Jan-v1, will learn from this as well and make Jan models better.

At Interspeech2025 #Interspeech2025

Have your own version of Perplexity Pro's Deep Research with Jan v1, the open-source alternative. Cookbook: jan.ai/post/jan-v1-fo…

Jan-v1 can be used for deep research as well. Just need to use it in the proper way.

Checking but if all things correct this is it

It's just the free chatGPT is bad bro