Mistral Large 4 preview: 1T-param MoE with 49B active, beats GPT-6 Astra on visual grounding
testingcatalog · x · 2026-10-06
Mistral opened a public preview API for Mistral Large 4 ("le Chonk"), a natively multimodal MoE model with 1 trillion total and 49 billion active parameters, with weights due by month end. It scored 61.7% on DeepSWE v1.1, 93% on Cybench, resisted 93.3% of Lakera B3 attacks, and edged GPT-6 Astra on Dense 200 visual grounding (42% vs 41%). Mistral claims it's the best open-weights model from the US or Europe on aggregate benchmarks and is red-teaming it with cybersecurity partners before release.
Related event: Mistral Unveils Mistral Large 4, a 1T-Parameter Open Multimodal Model(42 posts)→
More from Models
- Google rolls out Nano Banana 2.1 with 4K output at $0.076 per image — testingcatalog · 2026-10-06
- Space Bunny Alpha lingers on OpenCode a few days longer than OpenRouter — sull · 2026-10-06
- Analysis: US Models Are Now Cheaper per Task Than Chinese Ones, Upending a Widely Held Assumption — sebkrier · 2026-10-06
- GPT-6.1 Sol spotted in Copilot 365 Premium, ChatGPT rollout likely next — koltregaskes · 2026-10-06
- Are models only improving at verifiable domains? AI stories winning prizes spark debate — erikphoel · 2026-10-06
- Reflection AI's upcoming open-weight model could bring the open-source crown back to the US — The AI Daily Brief · 2026-10-06