They were given day zero access to release their quantizations[0]. The lowest quant will still require 256GB VRAM to run inferences, but as always I'm amazed by the Unsloth team by how much they can reduce the requirements for these powerful LLM's, and the easy-to-follow guides to get started with using them.
Unrelated to Large 4, but why is it so hard to give money to Mistral? Now they finally have a model that might be worth using, but I cannot for my life figure out how to give them money and be allowed to use their model. Its really shouldn't take more than a couple of clicks to get to a payment option.
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!
> 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.
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 :)
It's not open weight, it's weights available as they use a modified MIT license that is not OSI. It's the common "if you're serving it as a business you need a license" difference, for what it's worth.
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!
I work at a university data center and we host open weight models (as usual, this is my private opinion). I would consider open weight something you can take and self-host for free, also commercially and also that you can use for training and research. Anything else is just, I don't know, just "usable for free in certain cases"?
Open source or source available is something else entirely. To consider something source available I would require the actual training data to be available too. That is exceedingly rare.
"Open weight" is obviously intended to be analogous to "open source". The name only exists because people rightly complained about people calling weights "open source" when they were not. As such, the most reasonable interpretation is something that obeys the Open Source Definition as closely as possible, but ignores clause 2 because obeying it is legally impossible when the source code infringes copyright (as is the case with almost all LLMs). If it ignores any of the other clauses, it's instead closer to "source available" software, so the correct analogous name is "weights available".
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" ?
The OSI actually published an "open-source" definition for model weights, which requires the training code and training data composition to be published. The exact training corpus can remain private though.
"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.
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.
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.
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).
At the original pricing in my evals, Mistral Large 4 Preview was not competitive. With this price point and some improvements on top, this full release warrants a major reassessment.
based on Anthropic's S1 no one is serving models at a loss, but cutting your price to keep usage high makes a lot of sense when you have your inference booked for long periods.
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).
I just checked, seems you are right. They had roughly a week with the lower pricing, then increased it, now decrease it again. Very odd. My testing was with the higher pricing and looking at AA, theirs seems to have been too.
> 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.
> 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].
Finally, Le Chaton Fat. However, I'm not sure how well Mistral is doing in comparison to the other open-weight models, particularly Chinese ones, and I guess one would only really use Mistral if they want to have EU provenance for models.
"Cyber benchmarks, while achieving an average refusal rate of 96.6% on malicious cyber requests, the highest we measure among open models."
I feel like stating the obvious here and hence like I'm missing something. Where is the statistic about not refusing the non malicious requests? I can achieve the same performance with refusing when random(1000) < 967.
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.
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.
It's unfortunate, because the security theater performed here detracts from the model's overall cybersecurity performance. And uncensoring has another small hit to performance.
I have access to a server with 4 x h200 cards - 540 GB in total memory, but it appears I have to go to 4bit or 3bit even to serve it. Any idea if that is feasible/desirable (we have restrictions that we can't use a chinese model!).
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.
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.
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 reply 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.
Mistral has been great for us as a provider for our EU-based company. We use Mistral Large 4, GLM 5.3 Flash and DeepSeek v4.1 Flash 1017 through their API for our B2B2C app for complex work around building blocks for documents and custom tools. A few months ago we weren't sure which EU provider to use, but since they introduced 3rd party models, added more features (now Responses API etc.) and improved their own model, we are quite happy-
> 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.
So they can use GLM from a trusted provider, that is highly committed to respect local laws and regulations.
Just think about a defense contractor. Choosing trustworthy vendors is crucial to them. They can't just use a random provider from Open Router. They need to vet and certify their vendors. If they already have Mistral on the list of approved vendors, they can now just switch to the most recent open weight LLM if Mistral starts serving it.
There's many EU companies that want to have a contractual relationship with only another EU company and be absolutely assured that the models, the contents of the KV cache, any logging, all access control/credential systems, billing platform, etc are all hosted on EU-company owned hardware located in an EU jurisdiction.
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.
Their last 2 models were a disaster and didn't live up to their self reported evals. I hope this time around is much better. I'm happy tho that they are still figuring out how to get better even if only measured against themselves. This is great, the more options we have the better for everyone.
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, 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).
> 2. If the Licensee or any of its affiliates operates a Model as a Service business, the Licensee must enter into a separate commercial license with Mistral before using the Model or its derivative works for any commercial purpose. The Licensee may contact Mistral, using this form: https://mistral.ai/contact/, to request a commercial license.
Calling this open weight is misleading. That's why we had huge backlash when Redis, Elasticsearch, and Terraform went source-available, right?
Weight-available is a better adjective for this kind of licensing. Same for Qwen, GLM, Kimi, MiniMax, etc. If they really want to be open source, they should pick an OSI-approved license.
This is ancient (2 years ago), all their recent stuff (e.g. [0][1]) are MIT licensed. The only exception i've found from a quick browse of their stuff is [2], which is Apache 2 licensed. Even the update to V3 (which you linked at), V3.2, was MIT licensed.
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 restricting their next models to internal use? What about public chatbots?
Existing models 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.
There's always going to be some nerds out there that want to make and release a model as open, and share their knowledge along the way. I'm not too concerned.
The fear you're sharing has been shared since day 1 of open-weight models and it has yet to come to fruition.
Further, regardless of what any of these labs license their model as, the end user who has downloaded the weights can still use them for virtually any purpose (synthetic data, classification, etc) as much as they want and there is no way to reasonably enforce a license on those outputs.
So even if you can't run a service using this model's weights (without a license), the weights still do provide utility for training new models with any other license.
The problem is the massive amount of money, data, and talent to produce anything remotely competitive. Ideas have been floated about distributed training, but training requires insanely fast interconnect between nodes making it impossible with the current methodologies we have. Now if you mean funded at a national level, that seems more doable so long as the training runs are centralized in big DCs.
Great idea, let’s counter the power of a handful of large private organisations by putting that power in the hands of a single large government. What could possibly go wrong?
There are lots of governments around the world - some big, some small. Lots of discussions around digital sovereignty are framed in this matter - a model for Brazil, a model for Japan, etc.
We already had that world, the axis of evil stopped releasing open source models years ago. It’s only because of the Chinese that we kept getting them.
If it is capital intensive and we require the work be done in the public benefit, the only options are international consortia or nationalization. Non-profits will be guided by the interests of the rich and are unaccountable to the public. Worker cooperatives are unlikely to be well capitalized (because what rich person would want to invest and not have control) and will still have private profit incentives.
Peer production processes (like open source) require that the inputs be specialized knowledge and relatively small amounts of capital.
Of course they can. They're expensive to host, and if they're legally scrubbed off of places like huggingface then they are for all practical purposes 'closed' as in unavailable to people who don't already use them and if they're illegal or outpaced by Sota the ones who have copies can't even use them.
Thousands or more people already have copies and the license explicitly is irrevocable and allows redistribution. So if they take it off HF and HF is ok with that, people will just upload to anywhere else.
"outpaced by SOTA" doesn't matter at all, the models we have are plenty good enough for most real tasks.
A completely different direction is to remove the fair use allowance for training on copyrighted information (unless you e.g. open source your weights).
I'm a little confused by your distinction between vendor and lab. AI2 (Allen Institute for AI) is a western lab and they did release a fully open source model, with open weights, data, and training code (Olmo3).
Open source doesn't mean "anyone can use the source code commercially". Open weights similarly doesn't mean the weights can be used for anything. It just means the weights are open for anyone to download and inspect.
> Open source doesn't mean "anyone can use the source code commercially".
Yes it does. It might have requirements that you release the code for any changes you make to it or link with it. But no OSI (or FSF) approved license prohibits you from using it commercially.
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.
However, there is a more straightforward solution. The allows general use inside a company. The license blocks the "Model-as-a-Service".
So instead, any GPU holder can provide a 'Infrastructure-as-a-service' and have clients provide the model, and then tell the clients that as an optimization they can also provide just the hash and they'll load it from their memory.
I think its good that this license pushes the ecosystem towards such a generic-utility framework, but as a moat to protect a business its banking on most companies being lazy and wasteful.
Unless considerable amount of lobbying reverses it in a new directive/regulation, that's the case on model outputs - Originally I learnt of it from briefing prepared by lawyers regarding EU laws on AI.
It seems basically that LLM output is public domain, but a bit of human-generated content can be mixed in (whatever is enough to pass the "threshold of originality") and then the overall work can be copyrighted. Much like republishing a public domain book and adding an introduction or a few footnotes (but harder to extract the public domain component.)
But this all says nothing about the original question, about whether the weights can be copyrighted.
>> The emerging consensus across different jurisdictions is that purely AI‑generated content
with minimal to no human creativity does not qualify for copyright protection. By contrast, works
created with the use of AI‑applications may still enjoy copyright protection, where substantial human
creativity has been exercised. Whether this threshold is met must be assessed on a case‑by‑case
basis.
This makes sense, I believe it will eventually become the clear case in the US as well. However, it does leave open the case where a human carefully crafts a series of prompts to produce an creative work that qualifies for copyright.
Further, this does not appear to be binding on courts within the EU?
This doesn't seem like an actual court decision? Also it specifically says "if you use AI as a tool... then you probably own it" which, how else can you use AI? It doesn't do anything without a human wielding it, which has been the basis for all copyright claims so far.
Copyright doesn't care whether it's expensive. The fact that someone else, with the infrastructure and training data could get the same or almost the same weights makes it obvious there's no human originality in the weights, which means they aren't eligible for copyright. Some aspect of the infrastructure or processing of the training data might be copyrightable (or patentable), but the weights themselves? No human originality. No copyright.
I’m pretty sure they’re covered by the so-called sui generis protection of databases within the EU/EEA at least. That one is about protecting the effort and investment, not originality of expression.
"In the meaning of the European Union Directive 96/9/EC on the legal protection of databases,the term database refers to a collection of independent works, data or other materials, which have been arranged in a systematic or methodical way, and have been made individually accessible by electronic or other means. In the meaning of the Directive the data or materials:
must not be linked, or must be capable of separation without losing their informative content;
must be organised according to specific criteria, which means that only planned collections are covered;
must be individually accessible – mere storage of data is not covered by the term database."
It seems like model weights does not clear the definition of a database in EU law- notably, they are not a collection of individual works that are individually accessible.
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.
I would compare them more to media released under the creative commons license.
For example, you can record yourself playing different instruments, singing, and making other sounds. Then you can go combine all these recordings into a song, and you release the song under a CC license. The CC licenses framework does not demand that you also release the component recordings. But the released song can be remixed by someone else.
(Not making a value judgement, just making an equivalence).
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.
That's a strange take when it comes to AI companies who stole their training data in the first place. The AI companies sure wanted others to do things for them for free. What would "funding an effort" look like here? Bribing companies to share what they got from their stolen goods? Funding an effort to steal everyone's work too and then use it to train a truly open model ourselves?
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.
Open weights was already a weaker term then open source, but this is not even that. "Weights available" feels like an apt description akin to source available.
It doesn't really work like that. There are "keys" where you train in a specific response to a variety of queries. You can't really fine tune that away either.
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.
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 sustainable and sufficient that incentive can be.
Open weight models are kind of weird because they're usually just a bunch of weights and maybe a chat template, not even test vectors or anything. Quite odd, and probably one of the reasons why vllm, sglang, the llamas etc. so often produce buggy output. It's actually quite strange how little attention is given to verifying these setups work.
Why cant the community piece together the open-weights using open community-based like, implementation or whatever magic sauce integrates those weights
Someone’s surely doing it atm we just don’t know yet. And then we scan all books but the stuff stays local. Perhaps possible, perhaps not. But I don’t see why everyone could not run given the epoch on his content only, reporting back findings (loss).
It's not understandable for them to lie or mislead. Calling something open weights and not being remotely open in the commonly understood sense is misleading and not ok.
I don't think "you can do whatever you want other than just turn around and sell a chat wrapper for this" makes it immediately not open weights. Could it be more open? Sure. That doesn't mean it's misleading or a lie and I would strong disagree that it's "not being remotely open."
Their existence is the only thing keeping costs down across the board. If they had gotten open-weight model importing banned in the US a year ago, I'm sure they'd have 100% increased token cost relative to the advertised 100%+ intelligence improvements.
> Arguably hardware companies like Nvidia have incentives to subsidize open models to drive GPU demand.
This does not mean that they would necessarily subsidise consumer demand. They may be very happy to sell palette-loads of GPUs to datacenter and large corporations.
> Arguably hardware companies like Nvidia have incentives to subsidize open models to drive GPU demand
Not speaking on behalf of my employer and I have no insight into this strategy at all, but I feel like the sale of consumer GPUs is so small by comparison to data centers (more than 10x from what I can tell online) that more users of open weights wouldn't likely make much difference in the earnings report.
There’s no reason to believe that companies running open weights models would use consumer hardware. I can speak on behalf of my employer and say that we’d vastly prefer to have enterprise grade servers running inference on premises.
Not consumer demand. I was assuming that it's Nvidia's interest that intelligence is not only in a couple of frontier lab's hands, or they can choke the demand to suit their needs. If any company can access a frontier model without going through the frontier labs, the ecosystem will be larger and healthier for everyone, including Nvidia.
Open weights models in the data center are more important than on consumer devices.
In the future, even consumers will launch weights and fine tunes they own in the cloud.
People don't own the fiber to their home, so why should they expect to run the GPUs? If they want smart coding and video models, they'll need H200s and B200s, and they'll need to rent them.
counterpoint: I have smart coding at this very second, and I am running it on 128GB of ddr4 with a 16GB gpu. (Qwen 3.8 Flash Next).
I expect to continue to have this, and more, as smaller models become even more capable. I also expect hardware prices to resume going down once the supply crunch ends.
I see no need to stop owning things or change my expectations around ownership, nor will I accept something like that.
No, they're still capable of calling it something different, instead of implying open source licenses that they aren't actually using. That's a valid complaint that's kind of unrelated to any of this other stuff.
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.
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.
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.
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.
Contract law covers the use of licensed materials -- The license is the enforcement mechanism. Agreement of the model terms is implied by the use of the model.
If someone were to create a derivative model and distribute that, it would get murky if they licensed it Apache 2.0 or MIT, because users of that derivative would have never agreed to the upstream license (and may not even be aware of it), and the model is unlikely to be copyrightable.
Ughh the overly positive and cheerful comments here from what is usually the most critical/pessimistic audience on the internet is a huge tell.
If we cut through the political bias and wishful thinking for a second, here’s the truth: The model is not competitive. I would never waste my time using it for coding work, and I’m someone in the EU.
Great effort, but keep at it. This just isn’t good enough.
I’m sure you guys can make a boatload of money with regulatory capture and compliance games in the EU, but please don't give up trying to produce something thats actually competitive.
> 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?
It's nice they recently added acclaimed open models at a fair price like DeepSeek and GLM Flash, that way it's less of a gamble to try this new model out.
I added this to my weird little personal benchmark ( https://botbowlbench.com/) a couple of weeks ago, and it's...fine, but certainly not up to the best Chinese or US open weight models.
this "open weights" stuff is tantamount to shareware in the olden times... i hope it will soon be replaced by proper free software, where the actual training base is also given
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.
Who cares? It's a crappy model and not even truly open-weight. Mistral gets absolutely mogged by Deepseek and GLM. They have never been competitive and never will be. While China and US are competing for the olympic podium, Mistral is still in the kiddy pool.
At least the FLUX vision models from Europe are pretty good.
259 comments
[ 0.31 ms ] story [ 51.6 ms ] thread[0] https://unsloth.ai/docs/models/mistral-large-4
I have no idea how you could shrink this down to fit on any type of prosumer hardware though. I doubt it's possible.
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.
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.
This is on par with European stereotypes about French engineering so I’ll just go along with it :)
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!
Open source or source available is something else entirely. To consider something source available I would require the actual training data to be available too. That is exceedingly rare.
Also commercially? You think it doesn't count as open weights if you aren't given free reign to resell access to the model?
That's a very strange perspective to me.
They aren't saying you can't use it for commercial purposes. They're saying you can't sell access to it like an API without a specific license.
If it only restricted resale that would be okay in my book to call open weight. But this license is stifling.
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" ?
https://opensource.org/ai/open-source-ai-definition
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.
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.
At the original pricing in my evals, Mistral Large 4 Preview was not competitive. With this price point and some improvements on top, this full release warrants a major reassessment.
[0] https://docs.mistral.ai/inference/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).
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).
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.
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
Now, perhaps Kimi 3.1 is missing from the benchmarks, and Deepseek 4.1 Flash is still has impressive!
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.
I feel like stating the obvious here and hence like I'm missing something. Where is the statistic about not refusing the non malicious requests? I can achieve the same performance with refusing when random(1000) < 967.
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.
Guardrails on open-weight models is security theater purely to shield whoever released them from a potential PR nightmare.
"Our product won't hack you 96.6% of the time"
4 bit quants can be fine, try it and see.
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.
Just think about a defense contractor. Choosing trustworthy vendors is crucial to them. They can't just use a random provider from Open Router. They need to vet and certify their vendors. If they already have Mistral on the list of approved vendors, they can now just switch to the most recent open weight LLM if Mistral starts serving it.
I'm going to pass, but nice to know it's open now.
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, 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).
Calling this open weight is misleading. That's why we had huge backlash when Redis, Elasticsearch, and Terraform went source-available, right?
Weight-available is a better adjective for this kind of licensing. Same for Qwen, GLM, Kimi, MiniMax, etc. If they really want to be open source, they should pick an OSI-approved license.
Is there actually a competitive proper actually open model?
Correct me if I am wrong.
Latest versions of Kimi, Qwen, GLM aren't open, you need to pay them for commercial stuff and comply with PRC laws
[0] https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017
[1] https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813
[2] https://huggingface.co/deepseek-ai/DeepSeek-Math-V2
Existing models 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.
The fear you're sharing has been shared since day 1 of open-weight models and it has yet to come to fruition.
Further, regardless of what any of these labs license their model as, the end user who has downloaded the weights can still use them for virtually any purpose (synthetic data, classification, etc) as much as they want and there is no way to reasonably enforce a license on those outputs.
So even if you can't run a service using this model's weights (without a license), the weights still do provide utility for training new models with any other license.
Peer production processes (like open source) require that the inputs be specialized knowledge and relatively small amounts of capital.
"outpaced by SOTA" doesn't matter at all, the models we have are plenty good enough for most real tasks.
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.
No major Chinese or Western labs satisfy these requirements. The only vendors I can think of that do are Nvidia and Allen AI.
Their technical report on the release is excellent: https://allenai.org/blog/olmo3
They're a public research institute as opposed to one of the proprietary frontier labs, but they've been very open about the science.
Now we have these "open inspired but not open" licensed weights and they want to keep the term.
I'll stick with DS4.1 flash. Open and capable
Yes it does. It might have requirements that you release the code for any changes you make to it or link with it. But no OSI (or FSF) approved license prohibits you from using it commercially.
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.
That's not the way I remember it...
However, there is a more straightforward solution. The allows general use inside a company. The license blocks the "Model-as-a-Service".
So instead, any GPU holder can provide a 'Infrastructure-as-a-service' and have clients provide the model, and then tell the clients that as an optimization they can also provide just the hash and they'll load it from their memory.
I think its good that this license pushes the ecosystem towards such a generic-utility framework, but as a moat to protect a business its banking on most companies being lazy and wasteful.
That pays off more than people like to admit.
Unless considerable amount of lobbying reverses it in a new directive/regulation, that's the case on model outputs - Originally I learnt of it from briefing prepared by lawyers regarding EU laws on AI.
But this all says nothing about the original question, about whether the weights can be copyrighted.
This makes sense, I believe it will eventually become the clear case in the US as well. However, it does leave open the case where a human carefully crafts a series of prompts to produce an creative work that qualifies for copyright.
Further, this does not appear to be binding on courts within the EU?
https://intellectual-property-helpdesk.ec.europa.eu/regional...?
For example, you can record yourself playing different instruments, singing, and making other sounds. Then you can go combine all these recordings into a song, and you release the song under a CC license. The CC licenses framework does not demand that you also release the component recordings. But the released song can be remixed by someone else.
(Not making a value judgement, just making an equivalence).
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.
Source-available: source code available, but only for viewing, not to make derivatives.
Freeware/shareware: source code not available, but software is free to run.
Not sure what the weights equivalent of that last one is, but I feel like this is more like source-available.
Put a JEV on the output to filter out responses that might show its Mistral
Theft is the AI way
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.
And Hashicorp's software is no longer open source, after they changed the license.
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 sustainable and sufficient that incentive can be.
[0] https://www.forbes.com/sites/maryroeloffs/2026/08/17/ai-comp...
https://arxiv.org/abs/2603.08163
https://huggingface.co/1Covenant/Covenant-72B
Surely they'll outcompete everyone on cost alone
This does not mean that they would necessarily subsidise consumer demand. They may be very happy to sell palette-loads of GPUs to datacenter and large corporations.
Not speaking on behalf of my employer and I have no insight into this strategy at all, but I feel like the sale of consumer GPUs is so small by comparison to data centers (more than 10x from what I can tell online) that more users of open weights wouldn't likely make much difference in the earnings report.
In the future, even consumers will launch weights and fine tunes they own in the cloud.
People don't own the fiber to their home, so why should they expect to run the GPUs? If they want smart coding and video models, they'll need H200s and B200s, and they'll need to rent them.
I expect to continue to have this, and more, as smaller models become even more capable. I also expect hardware prices to resume going down once the supply crunch ends.
I see no need to stop owning things or change my expectations around ownership, nor will I accept something like that.
Call it free, available, downloadable etc, but not "Open".
Either way, Mistral is a failing company with little relevance in the AI space and that's not going to save them.
Kimi/GLM/etc basically all have this. They're not doing this out of the kindness of their own hearts.
https://commonsclause.com
permissionless vs permissioned licenses
many of us in open source ascribe to the permissionless ethos, but there is a spectrum or venn
It looks extremely clear on this page: https://docs.mistral.ai/models
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.
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.
In recent years many companies mislabeled source abailable code as open source as a marketing and to build a comunity before the rug pull.
Um. Yes. That is explicitly what open source means.
Inference can be a commoditized business.
When it comes to words, popular usage defines (and redefines) the meaning.
If someone were to create a derivative model and distribute that, it would get murky if they licensed it Apache 2.0 or MIT, because users of that derivative would have never agreed to the upstream license (and may not even be aware of it), and the model is unlikely to be copyrightable.
If we cut through the political bias and wishful thinking for a second, here’s the truth: The model is not competitive. I would never waste my time using it for coding work, and I’m someone in the EU.
Great effort, but keep at it. This just isn’t good enough.
I’m sure you guys can make a boatload of money with regulatory capture and compliance games in the EU, but please don't give up trying to produce something thats actually competitive.
Source: trust me bro
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?
At least the FLUX vision models from Europe are pretty good.