Mercury 3
The diffusion LLM that scaled 40x without slowing down
Proof1 independent outlet
Sources (1)
- [1]Mercury 3Product Hunt: AI launches · Oct 21, 09:51 AM
The diffusion LLM that scaled 40x without slowing down
Extractive summary: sentences quoted from the sources.
Before this
- Oct 21, 2026Flat-RLVR: Sharpness-Aware Reinforcement Learning with Verifiable Rewards for Reasoning Capacity in LLMs
- Oct 20, 2026Voxtral 3
- Oct 20, 2026[AINews] DeepSeek DeepInfer: Open Source SOTA Inference running 155T toks/day/cluster making $22m/day at ~80% margin for a 81 day payback
- Oct 20, 2026MiniMax H3 - Really is the best open-source video model.
- Oct 19, 2026llama.cpp support for Cohere models
- Oct 19, 2026Maxwell: A new kind of model that sits between code and hardware
