Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Mistral 3 family of models released

mistral.ai

Post reply on HN
Print view View original

171–180 of 243 posts

Re: Mistral 3 family of models released

#171

Post by sofixa »

BoorishBears wrote: post #165
sofixa wrote: post #162

Earlier quoted context omitted.

Mistral are mostly focusing on b2b, and for customers that want to self-host (banks and stuff). So their founders being from Meta, or where their cloud platform are hosted, are entirely irrelevant to the story.

The fact they would not exist without the leeches and built their business on the leeches is irrelevant. Pan-nationalism is a hell of a drug: a company that does not know you exist puts out an objectively awful release, and people take frank discussion of it as a personal slight.

> The fact they would not exist without the leeches and built their business on the leeches is irrelevant.

How so?

Re: Mistral 3 family of models released

#173

Post by GaggiX »

troyvit wrote: post #163
simonw wrote: post #121

The 3B vision model runs in the browser (after a 3GB model download). There's a very cool demo of that here: https://huggingface.co/spaces/mistralai/Ministral_3B_WebGPU Pelicans are OK but not earth-shattering: https://simonwillison.net/2025/Dec/2/introducing-mistral-3/

I'm reading this post and wondering what kind of crazy accessibility tools one could make. I think it's a little off the rails but imagine a tool that describes a web video for a blind user as it happens, not just the speech, but the actual action.

This is not local but Gemini models can process very long videos and provide description with timestamps if asked for.

https://ai.google.dev/gemini-api/docs/video-understanding#tr...

Re: Mistral 3 family of models released

#175

Post by cycomanic »

s_dev wrote: post #52

I was subscribing to these guys purely to support the EU tech scene. So I was on Pro for about 2 years while using ChatGPT and Claude. Went to actually use it, got a message saying that I missed a payment 8 months previously and thus wasn't allowed to use Pro despite having paid for Pro for the previous 8 months. The lady I contacted in support simply told me to pay the outstanding balance. You would think if you mis…

I'm not sure I understand you correctly, but it seems you had a subscription missed one payment some time ago, but now expect that your subscription works because the missed month was in the past and "you paid for this month"?

This sounds like the you expect your subscription to work as an on-demand service? It seems quite obvious that to be able to use a service you would need to be up to date on your payments, that would be no different in any other subscription/lease/rental agreement? Now Mistral might certainly look back at their records and see that you actually didn't use their service at all for the last few month and waive the missed payment. And that could be good customer service, but they might not even have record that you didn't use it, or at least those records would not be available to the billing department?

Re: Mistral 3 family of models released

#176

Post by troyvit »

cmrdporcupine wrote: post #127
mvkel wrote: post #93

Earlier quoted context omitted.

Open weight LLMs aren't supposed to "beat" closed models, and they never will. That isn’t their purpose. Their value is as a structural check on the power of proprietary systems; they guarantee a competitive floor. They’re essential to the ecosystem, but they’re not chasing SOTA.

This may be the case, but DeepSeek 3.2 is "good enough" that it competes well with Sonnet 4 -- maybe 4.5 -- for about 80% of my use cases, at a fraction of the cost. I feel we're only a year or two away from hitting a plateau with the frontier closed models having diminishing returns vs what's "open"

I think you're right, and I feel the same about Mistral. It's "good enough", super cheap, privacy friendly, and doesn't burn coal by the shovel-full. No need to pay through the nose for the SOTA models just to get wrapped into the same SaaS games that plague the rest of the industry.

Re: Mistral 3 family of models released

#177

Post by troyvit »

lalassu wrote: post #46

It's sad that they only compare to open weight models. I feel most users don't care much about OSS/not OSS. The value proposition is the quality of the generation for some use case. I guess it says a bit about the state of European AI

Glad I'm not most users. I'm down for 80% of the quality for an open weight model. Hell I've been using Linux for 25 years so I suppose I'm used to not-the-greatest-but-free.

Re: Mistral 3 family of models released

#178

Post by Legend2440 »

mbowcut2 wrote: post #125
barrell wrote: post #38

I use large language models in http://phrasing.app to format data I can retrieve in a consistent skimmable manner. I switched to mistral-3-medium-0525 a few months back after struggling to get gpt-5 to stop producing gibberish. It's been insanely fast, cheap, reliable, and follows formatting instructions to the letter. I was (and still am) super super impressed. Even if it does not hold up in benchmarks, it still out…

It makes me wonder about the gaps in evaluating LLMs by benchmarks. There almost certainly is overfitting happening which could degrade other use cases. "In practice" evaluation is what inspired the Chatbot Arena right? But then people realized that Chatbot arena over-prioritizes formatting, and maybe sycophancy(?). Makes you wonder what the best evaluation would be. We probably need lots more task-specific models. T…

I don’t think benchmark overfitting is as common as people think. Benchmark scores are highly correlated with the subjective “intelligence” of the model. So is pretraining loss.

The only exception I can think of is models trained on synthetic data like Phi.

Re: Mistral 3 family of models released

#179

Post by Jackson__ »

msp26 wrote: post #64

The new large model uses DeepseekV2 architecture. 0 mention on the page lol. It's a good thing that open source models use the best arch available. K2 does the same but at least mentions "Kimi K2 was designed to further scale up Moonlight, which employs an architecture similar to DeepSeek-V3". --- vllm/model_executor/models/mistral_large_3.py ``` from vllm.model_executor.models.deepseek_v2 import DeepseekV3ForCausalL…

So they spent all of their R&D to copy deepseek, leaving none for the singular novel added feature: vision.

To quote the hf page:

>Behind vision-first models in multimodal tasks: Mistral Large 3 can lag behind models optimized for vision tasks and use cases.

Post reply on HN