Mistral Large 4
funnym0nk3y
318 points
94 comments
October 21, 2026
https://docs.mistral.ai/models/mistral-large-4-0
Related Discussions
Found 5 related stories in 128.9ms across 10,209 title embeddings via pgvector HNSW
- Mistral Large 4 Philpax · 1677 pts · October 06, 2026 · 99% similar
- Mistral Large 4 apt-get · 70 pts · October 06, 2026 · 92% similar
- Mistral Large 4: "Le Chonk" j-bu · 518 pts · October 06, 2026 · 85% similar
- Mistral OCR 4.1 spelk · 294 pts · August 13, 2026 · 70% similar
- Mistral raises €3B to make sovereign, open-weight AI the technology frontier kuberwastaken · 14 pts · September 08, 2026 · 65% similar
Discussion Highlights (12 comments)
moecables
compared to Deepseek 1017, I'm not sure why I'd use this? https://openrouter.ai/compare/mistralai/mistral-large-4-0/de...
yorwba
Previous discussion: https://news.ycombinator.com/item?id=49977979 (2044 points 15 days ago, 1254 comments)
Flux159
I think that this link is more helpful: https://mistral.ai/news/mistral-large-4/ Seems like it's multimodal and visual intelligence capabilities are better than Deepseek 4.1 Flash, even though coding is not equal to DS 4.1 1017. The cost is about 10x of Deepseek though comparing on OpenRouter. Weights will be here https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B at the end of the month.
andix
I see Le Chaton Fat and press upvote, sure. But why? Wasn't it released already 2 weeks ago?
nik736
The weird thing about this model is that it spits out Chinese characters just like the chinese models when chatting with it in German. Happened to me quite often during the past few days.
dropbox_miner
waiting for pelican bench results (the only benchmark that matters)
maelito
What's more impressive is the datacenter used for training. It shows you can get "sub-SOTA" for one tenth of the resources of SOTA.
skohan
Impressive if these benches hold up to real world use. Not groundbreaking, but certainly in the game. Would love to see them release some smaller open-weight models.
codeulike
"Covered by GDPR" is a very big plus for me. Also very simple training opt-out.
jakswa
52B active parameters and my qwen-3.8-flash-next model running at home is supposedly competitive? maybe I gotta test it to believe it. https://artificialanalysis.ai/models/comparisons/mistral-lar...
snoutdata
It's quite fast
dom96
Has anything changed? I already ran it on my adversarial programming language benchmark[1] a few weeks ago and it performed pretty poorly (on par with Kimi K3). 1 - https://bench.killswitch-lang.org/