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.
It's gonna be interesting to see the final version, since this is a preview model still. I'm not completely sure as to how this works, but from what I gathered it will be further trained during this period, so it might improve a bit.
Since cost is so much higher than deepseek though it really seems like the biggest reason to use it would be "in the mistral subscription" for personal use or "because you need to abide by european data privacy laws" for companies.
Mostly I'm happy they're still in competition, as I mentioned already elsewhere.
Are they still in competition? Or are they a government sanctioned monopoly? I’m not sure that will be competitive. There are no market forces to do so but pride.
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.
Does this suggest Mistral is distilling Deepseek et. al.? Can distillation keep passing through with high fidelity, or is it lossy like copying a VHS tape?
Well, Models are hard to just distill down to a single benchmark number, or even a few across multiple different benchmark - they all behave and work slightly differently and you can find some who work or feel better for you and your use-case specifically.
But you're right, Mistral Large 4 seems a much more expensive but overall weaker model across the board.
The reason to use them would be, well, because they're...Mistral. They're in EU, france is large in nuclear energy so the energy cost of training has much less of an environmental impact, the data protection laws are very good in EU. This can really be a deciding factor for many EU companies, and I think it's good that they now have something which is at least semi-usable. (apparently)
It's also good that other options exist, even if they are worse. Deepseek is kind of like a large specter looming over every model release, because it's just SO cheap and good. That doesn't mean we should abandon every other model.
Still, the model is not really competitive on price and abilities alone. I'm mostly happy to see Mistral still in the fight.
I say this with a slight tongue in cheek, but Mistral does have a better option: they recently added GLM 5.3 to their platform (and it works rather well). Not trained in the EU, though, of course ;-)
Interestingly Mistral has been selling GLM5.3 inference hosted in Europe on their own hardware, for those who are OK using a "not developed in Europe" model but run on hardware located in and controlled by a European corporation.
> for those who are OK using a "not developed in Europe" model but run on hardware located in and controlled by a European corporation.
For the record this is also possible in Europe on Privatemode[1] which also gives you enclave-based processing. And access to GLM-5.3 Flash as well as GLM-5.3
Indeed, I see no difference between a chinese model hosted in Europe vs a Europe-grown model.
You could argue that potentially you can train a backdoor into a foreign-grown model (that has been demonstrated to be possible in toy examples), but I would highly doubt that this would be the case today, given the potential reputational damage if discovered.
Remember Stuxnet? An american worm that was completely benign, travelled on usb drives in hiding, and only activated when it guessed that it's on Iranian plants, and iirc changed very subtly some very specific parameters of operations.
Here, with a chinese model, it would be even simpler - you finetune a model to activate certain operations when it discovers certain environments, but to behave completely normal otherwise. Kind of like a sleeper agent. And with current tech we wouldn't even be able to spot this.
I recall seeing something about that. I don't know of anyone still using such an ancient model when much newer Qwen 3.5-120, 3.6-35B-A3B, 3.8-27B and 3.8-flash-next all exist and score much higher on all conceivable metrics. I would like to see if anyone can replicate the same thing on a 'current' Qwen released in the past 6 months.
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.
It's gonna be interesting to see the final version, since this is a preview model still. I'm not completely sure as to how this works, but from what I gathered it will be further trained during this period, so it might improve a bit.
Since cost is so much higher than deepseek though it really seems like the biggest reason to use it would be "in the mistral subscription" for personal use or "because you need to abide by european data privacy laws" for companies.
Mostly I'm happy they're still in competition, as I mentioned already elsewhere.
Are they still in competition? Or are they a government sanctioned monopoly? I’m not sure that will be competitive. There are no market forces to do so but pride.
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...
Previous discussion: https://news.ycombinator.com/item?id=49977979 (2044 points 15 days ago, 1254 comments)
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.
Does this suggest Mistral is distilling Deepseek et. al.? Can distillation keep passing through with high fidelity, or is it lossy like copying a VHS tape?
I see Le Chaton Fat and press upvote, sure. But why? Wasn't it released already 2 weeks ago?
Some details and benchmarks were announced but the weights didn’t get released
Still 9 days before the release of weights apparently.
Le Chaton Fat is a lovely name. I predict the next Michelin restaurant with will have an LLM as the Chef...
compared to Deepseek 1017, I'm not sure why I'd use this?
https://openrouter.ai/compare/mistralai/mistral-large-4-0/de...
Well, Models are hard to just distill down to a single benchmark number, or even a few across multiple different benchmark - they all behave and work slightly differently and you can find some who work or feel better for you and your use-case specifically.
But you're right, Mistral Large 4 seems a much more expensive but overall weaker model across the board.
The reason to use them would be, well, because they're...Mistral. They're in EU, france is large in nuclear energy so the energy cost of training has much less of an environmental impact, the data protection laws are very good in EU. This can really be a deciding factor for many EU companies, and I think it's good that they now have something which is at least semi-usable. (apparently)
It's also good that other options exist, even if they are worse. Deepseek is kind of like a large specter looming over every model release, because it's just SO cheap and good. That doesn't mean we should abandon every other model.
Still, the model is not really competitive on price and abilities alone. I'm mostly happy to see Mistral still in the fight.
I say this with a slight tongue in cheek, but Mistral does have a better option: they recently added GLM 5.3 to their platform (and it works rather well). Not trained in the EU, though, of course ;-)
Yeah, absolutely true. And I am guilty of simply using Mistral as a GLM subscription myself. I just really like the GLM 5.3 model.
I will see where Chonk is better though, since you have both options in one subscription.
If you're required to use a European model
Interestingly Mistral has been selling GLM5.3 inference hosted in Europe on their own hardware, for those who are OK using a "not developed in Europe" model but run on hardware located in and controlled by a European corporation.
https://docs.mistral.ai/models/zai-glm-5-3
> for those who are OK using a "not developed in Europe" model but run on hardware located in and controlled by a European corporation.
For the record this is also possible in Europe on Privatemode[1] which also gives you enclave-based processing. And access to GLM-5.3 Flash as well as GLM-5.3
[1]https://www.privatemode.ai/
Indeed, I see no difference between a chinese model hosted in Europe vs a Europe-grown model.
You could argue that potentially you can train a backdoor into a foreign-grown model (that has been demonstrated to be possible in toy examples), but I would highly doubt that this would be the case today, given the potential reputational damage if discovered.
Remember Stuxnet? An american worm that was completely benign, travelled on usb drives in hiding, and only activated when it guessed that it's on Iranian plants, and iirc changed very subtly some very specific parameters of operations.
Here, with a chinese model, it would be even simpler - you finetune a model to activate certain operations when it discovers certain environments, but to behave completely normal otherwise. Kind of like a sleeper agent. And with current tech we wouldn't even be able to spot this.
Didn't we get a submission today about Qwen3-Coder producing more bugs when hinted on the user having different ideology than the Chinese government?
Ok I see it's flagged now, probably fake, but the fact that it's possible does make me feel uneasy about both Chinese and US models.
It was so insignificant, it could be equally described by the non-determinism inherent in the models. And that it reads more like a marketing piece.
That's my read anyhow. It's been flagged. The whole thing is kind of rich, considering the Israeli roots of the company.
https://news.ycombinator.com/item?id=50179706
I recall seeing something about that. I don't know of anyone still using such an ancient model when much newer Qwen 3.5-120, 3.6-35B-A3B, 3.8-27B and 3.8-flash-next all exist and score much higher on all conceivable metrics. I would like to see if anyone can replicate the same thing on a 'current' Qwen released in the past 6 months.
> I see no difference between a chinese model hosted in Europe vs a Europe-grown model.
Really? Fascinating. Do you know about Tiananmen Square, the great firewall of China, etc?
if you care about data-retention policy or you have to use european models ...
Patriotism.
Well when your greatest ally acts like an adversary, who can blame them for the increased sovereign focus?
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.
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.
"Covered by GDPR" is a very big plus for me. Also very simple training opt-out.
waiting for pelican bench results (the only benchmark that matters)
NVM. Looks like it already exists:
https://tools.simonwillison.net/markdown-svg-renderer?url=ht...