LaunchAI ModelsOctober 6, 2026

Mistral releases Mistral Large 4 (Le Chonk), a 1.05T-parameter MoE

Read original source →marktechpost.com

Mistral Large 4 is a granular MoE with 1.05T total parameters, 49B active per token, a 1.6B vision encoder and 1M-token context, trained from scratch on 3,800 NVIDIA Grace Blackwell GPUs in EU datacenters. API is live at $1.36/M input and $4.18/M output; weights ship end of October.

1 source

More stories today

Open the live feed