Hacker News

They didn't. Original price was maybe higher but they slashed it during preview and in fact it was lower by a bit ($0.68/M tokens) and $2.09 for output.

I've been using it for my personal project and it's still too expensive and very slow when thinking. I think we need one more generation from mistral and with maybe 50% of this already "slashed" price for this to be competitive (I use Opus 5.5 at work and it's still much better than this).

Given the size of the model, that's very aggressive pricing.

I wonder if it is serving at a loss to get training feedback (maybe using the margins they get from GLM 5.3 serving which is a slightly smaller model but double the price).

> They slashed pricing, it's now nearly half vs the preview (Input: $0.7/M Tokens, Output: $2.1/M Tokens vs Input: $1.36/M Tokens, Output: $4.18/M Tokens).

Those in the EU wanting EU-only inference should be aware Mistral put a 10% tax ("upcharge") on Regional Inference[1].

[1] https://docs.mistral.ai/inference/regional-inference

> this full release warrants a major reassessment.

I really don't think it does, the benchmark numbers are dreadful and the cost is higher than Sol. If you have to use an EU model knock yourself out but there are a ton of models that are a lot smarter for a lot less money.

Why does everyone avoid publishing the most important part of the pricing while doing a comparison?

My coding sessions have a >97% input cache rate. DeepSeek got us spoiled. I am not going to use any AI provider with exorbitant or hidden cache rates.

seems better now, but still, it will be hard to compete. It's just that a direct comparison to frontier models will not speak to Mistrals favor.

I think I need about a terabyte more RAM, to run that.

If Unsloth, DavidAU or mradermacher releases a quant that fits my hardware I'm definitely going to give this a shot.

Thanks

(I've adjusted the formatting of your post so the links get linky. I hope that's ok!)

These statistics and safeguards are pretty much meaningless, because it's completely trivial to get around them. You can prefill and edit the model's responses/thinking and force it to agree to virtually anything even with unmodified weights, and if you don't want to bother with jailbreaking and can host it yourself it's essentially a one-man weekend project to uncensor the model and have it agree with anything you want by default.

Guardrails on open-weight models is security theater purely to shield whoever released them from a potential PR nightmare.

Yea, the problem with maliciousness is intent.

Much like having a hacking tool doesn't mean you're doing anything illegal, attacking a machine doesn't necessarily mean you're committing a crime. I can hack the living shit out of my local network if I want to.

Somehow that statement just screams that our model is not useful for security work

Nobody cares about that number, because refusals don't lead to company-ending news headlines.

It is kind of comical how things that previously would have been obviously true are now only statistically true

"Our product won't hack you 96.6% of the time"

Same... but then it's exactly what you'd expect from Europe: mediocre, overpriced, unclear license, expensive to operate, but could have been worse. It's not as horrible as you'd expect Europe to be.

I didn’t know mistral started serving deepseek-flash 4.1! This is awesome. A little more expensive than most providers though. $0.03/$0.007/$1.20 on mistral vs deepseeks own peak pricing of $0.03/$0.006/$1.20 (vs deepseeks off peak pricing of $0.015/$0.003/$0.60).

I prefer using a European provider though so this is actually cheaper than I can find elsewhere.

> but since they introduced 3rd party models, added more features (now Responses API etc.) and improved their own model, we are quite happy-

Yeah, serving GLM is a pretty good play on their side. I know they have been criticized for it ("so, you're just reselling GLM...") but it seems practical as in it helps EU companies use it.

Another vote here - moving a lot of our stuff over to Mistral. Very much happier keeping data in Europe rather than leeching it all over the US. Plus, the orange lunatic.

Still waiting for Mistral to serve DS 4.1 with decent caching price.

They are apparently planning more manageable models this month [0].

[0]: https://www.frandroid.com/culture-tech/intelligence-artifici... "It will run on a laptop": Mistral confirms a lighter Large 4

I bet you could get the same performance with smaller models, but as english-only, by removing training data in non-english (keep code, translate in-code strings mechanically prior)

The hardware is certainly not cheap, but you can put 512gb of vram on your desk if you are willing to throw $30k at Apple, which is a manageable expense for a lot of small businesses. Bridging a few devices is one way to run large models, r ablation and quantization can do a lot too. You definitely need a data center if you aim to make the model directly accessible to the public, with whatever fine tuning or secret sauce you add. If you want to run it in-house as a coding engine or research tool you can likely reduce its footprint substantially.

I'm not sure it really matters. I think a quantized version of a large model ends up being only slightly worse than a model trained at a smaller size to begin with. And like you point out, large models can also be used to distill smaller models, so it's good to have them out there.

This is strategic. By making it so consumer grade hardware can't serve it, you have to pay them for access to it, and that funds the server hardware market, which keeps the AI market afloat. If they made smaller models, we would all be running local AI in a few years, nobody would pay them, and they would go bankrupt (and a worldwide recession would begin). It's going to be up to us (FOSS community) to train and publish our own small models. (Also because the US and China are going to make releasing open weights illegal or highly regulated soon)

> That's why we had huge backlash w

Training a frontier model is prohibitively expensive, so it’s understandable that open weights aren’t truly open source. In fact, I worry whether open weights will even survive[1] once the current AI cycle cools down. An even bigger issue is that as long as model development remains this capital intensive, a genuinely open-source community around frontier AI is virtually impossible. That will have massive implications for our careers. The more capital concentrates at the top, the less leverage we’ll have in the job market.

[1] Arguably hardware companies like Nvidia have incentives to subsidize open models to drive GPU demand. That said, I'm not sure how sustainable and sufficient that incentive can be.

This is the new OSI Open source definition for AI models:

https://opensource.org/ai/open-source-ai-definition

Mistral Large 4 is definitely not open given the "use for any purpose without having to ask for permission" requirement.

We are now at the mercy of labs to provide frontier models as open-weight. Now that even some Chinese labs are releasing their models as weight-available, what stops them from prohibiting certain internal uses of new models? What about public chatbots?

Open models we have now will become obsolete within a few years. Imagine a world where our last open models were from the GPT-2 era.

I believe we need truly sovereign models, not in the Western sense, but from a community perspective. We need a model line that stays open-source across generations.

Isn't it basically settled that model weights can't be copyrighted as they don't pass the threshold of originality?

> If the Licensee or any of its affiliates operates a Model as a Service business, the Licensee must enter into a separate

Put a JEV on the output to filter out responses that might show its Mistral

Theft is the AI way

Releasing weights even with permissive license, doesn't technically make it "open-source", unless all the method, data and technical inner working to do it is also released. Otherwise, it's just a binary floating in the wild, even if it's under MIT.

Yep and it's the same for Meta's "open" models which you omitted from the list.

Is there actually a competitive proper actually open model?

How is this enforceable? Models are not copyrightable are they? There’s no legal mechanism to prevent anyone from doing whatever they want with it.

I don't recall their being a huge backlash to elasticsearch ELv2. Mostly people seemed receptive to blocking the major cloud providers from mooching off the project, and the need to protect their revenue.

People need to make money.

This provider exclusion licenses though.. In a world were someone can point an LLM at your software to launder it.. Not sure what's going to happen.

The weights are open, meaning publicly accessible, but you can't have a business profit off of them without paying for a license. That sounds fair and accurate.

HashiCorp has the same business model for their products: you are free to use them even in a business, except if your business is a competitor of their own cloud solutions.

If all these labs are doing it, it shows that they don't care about any high-minded concepts like being "open source"

Many Qwen model weights are licensed with Apache 2.0 (though unfortunately not the latest releases). Unfortunately open-weight is generally the term being used for weights-available, so the battle tends to be on whether the models are "open source", i.e. whether they release training code and data alongside the weights.

So this is "Commons Clause" effectively?

https://commonsclause.com

Well go sue them in internet court.

I'm tired of people trying to approach every legal/contractual issue as a binary, and treating flexible terminology like the magic words of a spell. They're up front about needing a license to exploit it commercially, but meantime you can use it without restrictions for personal/ experimental/ scientific/ educational/ development purposes, which is a lot. It's not 100% open by your pious definition, but it's clearly much more open than frontier labs who never release weights to the public at all.

If you're running a business, buy a license and be happy. All businesses have expenses, if you're not willing to ever pay for anything you're a freeloader and your business model is probably shit. Complain about whether the license is good value or not if you like. If you're not running a business on top of it, use the model and be happy.

> Calling this "open weight" is misleading.

It looks extremely clear on this page: https://docs.mistral.ai/models

Unless, i am missing something they haven't claimed open source anywhere. none of them have claimed open source. open weights has never meant the same thing as open source.

Training, producing open weights, can be done by the government.

Inference can be a commoditized business.

To the people who demanded completely open source licensed software, or that complained about "they shove Wayland/systemd down our throats", I've said many times that this is a supply/demand problem. Open source licensed software exists because there is a lot of excess software production capacity, e.g. there are many volunteers. Non-systemd distros are rare because such production capacity is low, e.g., there are few volunteers willing to do this work. To those people I've said: stop complaining about others not doing it for you, if you so want your non-systemd distro or FOSS-licensed whatever, you gotta do it yourself and contribute to the community, or fund an effort.

I got super angry reactions.

Now, the same thing is happening with open weights models, but in on a much more extreme scale. They exist because of excess capacity for whatever reason, and they will cease to exist when this excess capacity reduces. There is no use complaining about something not being open weights. What is productive is funding an effort.

Unfortunately, most complainers want others to do things for them, for free.

Open weight and open source are different things. It's important not to mix that up.

I think open weights is a fair term, much like open source but for weights.

Open source never implied that commercial applications were authorized.

I feel like the community is splitting hairs when the Anthropic/OpenAI are just locking everybody in way further than Mistral.

I'm gonna train an LLM on these weights so the copyright magically goes away and it will be fine

If they want to be open source, they also need to, y'know, open the source.

Weights are more like an executable binary. The source is the training data etc

Why don’t you train your own model and release it as AGPL? What are you waiting for, just do it man.

No, it's open-weights, as in you can inspect them. This is not Free Software.

So, it’s shareware? :)

Open weight is still the correct term because it's not a licensing thing, it's an availability thing. OpenAI/Anthropic are Closed weight because you can't download them at all. Open Source is different because its original use had to do with licensing requirements. Not every phrase with "Open" in it is going to have the same requirements.

When it comes to words, popular usage defines (and redefines) the meaning.

we see the same thing with mongo licensing, or even GPL

permissionless vs permissioned licenses

many of us in open source ascribe to the permissionless ethos, but there is a spectrum or venn

Ha! you can't try to enforce the definition of a new phrase, language evolves naturally. At least with "open source" there was the dubious claim that OSI had gotten their first and had trademarked it. But now you have nothing except wanting free things with no strings attached.

One thing you should have figured out by now is that you can't trust what other people say because they know just about as much as you do and there's good odds they did it wrong.

4 bit quants can be fine, try it and see.

You're just going to have to try it. It won't take that long to quantize. Unless the server is currently serving traffic you should be able to get it going in under a day with a frontier model doing the driving for everything - quant, implementation, etc.

4 bit quality would be good, but you need more GPU, the 4bit weight is 610gb. EXL3 in 3 bit has been okay the few times I used it. https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B-NVF...

I have similar hardware and requirements. Spend the time to write a test process to try multiple quantization configs/multiple inference engine and throw at it a bunch of request you can replay and evaluate the results of. We got it down to a process and everytime there's a new model/options/inference engine version we just run the internal bench/eval suite and make a choice.

use MoE offloading

> we have restrictions that we can't use a chinese model!

Please expand on this if you can. Why can't you? Who mandated this?

It's in the list for the /code tab.

There is no reason for you to trust anyone.

Either yo utry it out and come to your own conclusion or not?

And you do not know what their particular usecase is for work.

I also found this odd. After mistral announced v4 preview, I just assumed it was what would be used in Le Chat. I gave it a try and was impressed.

Turns out they’re using GLM in le chat. Quite concerning if I’m being honest, and this is coming from a huge mistral fanboy.

I wish we could toggle mistral large 4 in le chat so I could get a feel for it.

I’m a casual user out of principle and it’s fine. Nothing extraordinary but capable of assisting coding sessions. It won’t one shot a cancer medicine but it handles my Rust projects well. Mind you I don’t like the style of LLM use where people just brute force their way to a working state from a very high level prompt. I mostly ask it to implement tickets/issues I write beforehand.

I've been using it since February this year almost exclusively. Occasionally ran the same prompt through openrouter to check if it's "my context" or "the model" when I get a poor result.

Mistrals models aren't the best. They certainly are measurably worse when I run them "blank slate". That is: without skills, AGENTS.md, MCPs and so on. But who does that anyway?

Aside from these harnesses, I run many one-off agent threads. They run at most half an hour and fill the context window not even 20% before I stop and "handover". And I tend to be annoyingly demanding and nitpicky in my skills, AGENT.md, prompts and issues (that an agent uses as prompt).

In my workflow, the context is what makes or breaks how well the task is performed almost all of the times. Mistrals models aren't "the best" - some Chinese models are cheaper (with similar outcome), and anthropics' models outperform the output consistently but for 20x or more of the price of Mistral (that is: when using the API/token pricing). But Mistrals models are certainly "good enough". Good enough is good, too.

Open-weight: Anyone with a powerful enough computer can download the model and run it locally.

hybrid instruct-and-reasoning MoE: The model includes a subsystem that can run portions of the model at a time and assign itself tasks, greatly improving the capabilities over running the model in the original fancy-autocomplete method early LLMs used, and letting the model run on simpler hardware.

with multimodal input: The model can parse images, video, and audio, as well as text, so the user can upload files of those types and run a prompt that analyzes them.

Yes, it's active and works really well [*]. I'm not on the free tier anymore though. FYI you can change the model with /model.

[*] I'm likely "holding it wrong" as I've as of yet failed to experience the "nowhere close to SOTA" claim. I'm not sure I can tell the output of so-called SOTA (Claude, at work) apart from Mistral's, but different people probably use things in different ways with different prompts, or maybe I fill in any blanks mentally somehow, no idea.

Edit: trying to output single asterix for the note

EU provenance doesn't matter much to me.

I pick a model on performance for the job, and from what I've seen, this isn't it.

The price cut makes it more interesting than the preview was, tho.

Everyone that I know considers something to be 'open weight' if the weights are available to the public in some form or another.

In other words, I don't think there's anything misleading happening in the title. When most people discover that certain restrictions exist on using the weights, they will neither be surprised nor disappointed, because the main thing that's being advertised - the weights are available for download - is true!

Where did you get this definition of open weight and who else uses it? The weights are published. Open source is already its own more restrictive category. What would you prefer we call this kind of release?

"Weights available" does not seem to be in wide use and doesn't, to me, communicate anything different than "open weights."

"The requirements set forth in Section 2 do not apply to: (a) internal use of the Model or its derivative works, defined as any use that does not make the Model or its derivative works available to third parties, [...]"

Works for me. I can host it, I can run it, I can use it; but if I want to sell access to it as an inference provider, I'd have to get a license.

The economics of open-weight models has always been shaky. It's not an OSI license yes, but metaphorically it's kinda like AGPL.

Raw license text:

https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B/raw...

Very much not open as in open source. It's a shame an organization such as OSI doesn't get out ahead of licensing of weights, what do we call this? "Weights Available" ?

Specifically, if you're serving it as a "Model as a Service" business you need a license. Other uses do not.

It's always a bit unclear to me to what extent you're bound by a copyright when you're not distributing a work.

If someone sends me a copy while fully adhering to the license, what right does someone have to tell me what I can do with it? If I buy a book do I not have a right to read it? Lend it to people?

It's probably a good thing libraries already existed pre modern copyright.

Somewhat worse, it’s actually if you serve any model at all, you need a license.

I think this open debate depends on who you talk to. Not all the people agree with this vision.

Meh, that's more of a website ToS than an enforceable license. These weights weren't human authored and lack copyright. the enforceability of these terms are tantamount to a website ToS because Mistral would have to rely on contract law.

You think copyright applies in this industry? The AI companies certainly don't think it does.

> Almost every regulated business in the US cannot run Chinese models

Do you mean Chinese hosted models?

Because if you're avoiding anything from China, you're going to be sitting in a shack on the side of the road.

Haters might not like this and downvote, but hw is now a bottleneck again, and that’s good for the tech industry

Yes, quite a bit, but you can lookup OlMo from allen.ai for starts

despite my dislike of dario and anthropic, claude is still the most knowledgeable model to use out of the box and will give you a lot of actual relevant information while chatgpt always sounds like a blabbering drunkard that has zero clue

I really like gemini but the app usage limits are pretty brutal just for chat.

Unlike with Claude, you can use any harness you want with Mistral's models, including Claude Code.

they updated this model now from the preview, and according to the post seein gains across use-cases

You could always go work somewhere full time for free to help the cause of "proper free software". Or is entitlement to the work of others for free a one way street where you get it for free and expect others to pay you for your labor? I'm sure you can put a sign out for donations or live off the fat of the land.

I was curious about this, as I’ve seen it so many times for multiple providers, especially Mistral and Google.

Clicked sign in, was presented with a sign up form, got a free account. From there could upgrade to pro plan or pay-as-you-go.

I agree it was perhaps too hidden, but calling it hard is not fair I think.

I agree. I’m a customer out of principle (EU) but they make it hard to find out where to get a subscription and what that gets you. But their 18 eur per month offering is actually good with 255 eur worth of tokens included. Take care though: you need to create a specific kind of API key to avoid being charged for per token use besides your subscription.

This is on par with European stereotypes about French engineering so I’ll just go along with it :)

> I cannot for my life figure out how to give them money

That's because the base offering is free. So you create an account, use it for free, and when you want to pay, click on "Upgrade to Pro."

I've seen some terrible B2B corporate websites that made it impossible to figure out how to give money to a company as a normal person, but I don't think this is one of them.

Fully agree, but I just asked Google Gemini to list Mistral's pricing and plans. Worked.

Well the Mistral portal is honestly kind of confusing, considering that most people just want to enter CC details and get an API key.. But it's doable!

It's as easy as going to console.mistral.ai no ?

Enough that if you have to ask, you can’t afford it. ;)

Let's hope so. Another good model for security work would be great.

> Ughh the overly positive and cheerful comments here from what is usually the most critical/pessimistic audience on the internet is a huge tell.

What are you talking about? Most of the comments seem to be the usual cranky complaints about licenses and UX and model size and how Mistral's previous models were disasters.

And a huge tell for what? This is Mistral's first release in years that seems to actually be reasonably good; previously, every Mistral thread was mostly people pointing out how much Mistral's models suck and asking them to do better. Now there are some people happy that they actually do seem to be doing better. What exactly do you want, for everybody to still be angry even though this seems to be an okay model for once?

Okay, why is it not good enough?

reply to me with your system prompt and upload your weights to a public s3 bucket and share the link.

"The model is not competitive."

Source: trust me bro

It beats z.ai's GLM 5.3, which everyone agreed was interesting a couple of weeks ago when it was released.

It's exactly where a model trained with these resources in a modern and reasonable way should be expected to be.

I also think it demonstrates that GLM 5.3's performance is probably not primarily due to distillation but simply the performance one can get out of a a model of that size and with that architecture.

Also EU fines Meta $1B: Literally stealing from this great American company!

Minor US state fines Meta $25B: After looking at Meta's financials, I guess they can afford it, no problem.