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

221–230 of 243 posts

Re: Mistral 3 family of models released

#221

Post by leobg »

bildung wrote: post #74
constantcrying wrote: post #6

Earlier quoted context omitted.

The lack of the comparison (which absolutely was done), tells you exactly what you need to know.

I think people from the US often aren't aware how many companies from the EU simply won't risk losing their data to the providers you have in mind, OpenAI, Anthropic and Google. They simply are no option at all. The company I work for for example, a mid-sized tech business, currently investigates their local hosting options for LLMs. So Mistral certainly will be an option, among the Qwen familiy and Deepseek. Mistral…

Does your company use Microsoft Teams?

Re: Mistral 3 family of models released

#223

Post by mortsnort »

I use a small model as a chatbot of sorts in a game I'm making. I was hoping the 3b could replace qwen 4b, but it's far worse at following instructions and providing entertaining content. I suppose this is expected given smaller size and their own benchmarks that show Qwen beating it at instruct.

Re: Mistral 3 family of models released

#224

Post by accrual »

Congrats on the release, Mistral team!

I haven't used Mistral much until today but am impressed. I normally use Gemma 3 27B locally, but after regenerating some responses with Mistral 3 14B, the output quality is very similar despite generating much faster on my hardware.

The vision aspect also worked fine, and actually was slightly better on the same inputs versus qwen3 VL 8B.

All in all impressive small dense model, looking forward to using it more.

Re: Mistral 3 family of models released

#226

Post by astrange »

pembrook wrote: post #160
mbowcut2 wrote: post #125

Earlier quoted context omitted.

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…

If the models from the big US labs are being overfit to benchmarks, than we also need to account for HN commenters overfitting positive evaluations to Chinese or European models based on their political biases (US big tech = default bad, anything European = default good). Also, we should be aware of people cynically playing into that bias to try to advertise their app, like OP who has managed to spam a link in the fi…

Americans have an opposing bias via the phenomenon of "safe edgy", where for obvious reasons they're uncomfortable with being biased towards anyone who looks like a US minority, and redirect all that energy towards being racist to the French. So it's all balanced.

Re: Mistral 3 family of models released

#227

Post by rc1 »

usrnm wrote: post #26
ot wrote: post #20

Earlier quoted context omitted.

Is it so hard for people to understand that Europe is a continent, EU is a federation of European countries, and the two are not the same?

Europe isn't even a continent and has no real definition (none that would make any sense, anyway), so the whole thing is confusing by design

If Europe isn’t a continent, on what continent are the EU member states sitting on?

Re: Mistral 3 family of models released

#228

Post by rkomorn »

rc1 wrote: post #227
usrnm wrote: post #26

Earlier quoted context omitted.

Europe isn't even a continent and has no real definition (none that would make any sense, anyway), so the whole thing is confusing by design

If Europe isn’t a continent, on what continent are the EU member states sitting on?

Eurasia is the widely accepted answer.

Re: Mistral 3 family of models released

#229

Post by distalx »

barrell wrote: post #79
mrtksn wrote: post #69

Earlier quoted context omitted.

Some time ago I canceled all my paid subscriptions to chatbots because they are interchangeable so I just rotate between Grok, ChatGPT, Gemini, Deepseek and Mistral. On the API side of things my experience is that the model behaving as expected is the greatest feature. There I also switched to Openrouter instead of paying directly so I can use whatever model fits best. The recent buzz about ad-based chatbot services…

Yep I spent 3 days optimizing my prompt trying to get gpt-5 to work. Tried a bunch of different models (some Azure some OpenRouter) and got a better success rate with several others without any tailoring of the prompt. Was really plug and play. There are still small nuances to each one, but compared to a year ago prompts are much more portable

What tools or process do you use to optimize your prompts?

Re: Mistral 3 family of models released

#230

Post by druskacik »

leobg wrote: post #210
druskacik wrote: post #94

Earlier quoted context omitted.

This is my experience as well. Mistral models may not be the best according to benchmarks and I don't use them for personal chats or coding, but for simple tasks with pre-defined scope (such as categorization, summarization, etc.) they are the option I choose. I use mistral-small with batch API and it's probably the best cost-efficient option out there.

Did you compare it to gemini-2.0-flash-lite?

I did some vibe-evals only and it seemed slightly worse for my use case, so I didn't change it.
Post reply on HN