Re:Mistral Large 4: "Le Chonk"
by booty on Tuesday October 06, 2026 @11:11AM
(#49979693)
Sans context, I really like the Anthropic and Claude's faux-academic, minimal-ish, intellectual-ish branding.
(Especially the way it looked ~12 months ago -- it's gotten more cluttered since then. Perhaps unavoidably, as the breadth of their offerings has grown)
But over time it's begun to feel like unsettling cognitive dissonance as their ambitions grow and the stuff to worry about has piled up.
Re:Mistral Large 4: "Le Chonk"
by monocularvision on Tuesday October 06, 2026 @12:40PM
(#49980932)
I know the whole “AI logos look like buttholes” thing is a joke for most people, but I really speaks to the pornification of our society. I would never have thought “butthole” looking at any of their logos.
Re:Mistral Large 4: "Le Chonk"
by walrus01 on Tuesday October 06, 2026 @10:36AM
(#49979182)
I think there's such a thing as throwing too many marketing and sales people and too much polishing and "refinement" at something. I see in new product announcements from Microsoft as well. It's like seeing someone try too hard to impress you.
Re:Mistral Large 4: "Le Chonk"
by YeahThisIsMe on Tuesday October 06, 2026 @11:46AM
(#49980234)
I do my best to avoid any AI marketing because I extremely despise it.
I just haven't quite decided on my new profession, yet, but it's either going to be with plants or with animals.
Re:Mistral Large 4: "Le Chonk"
by vessenes on Tuesday October 06, 2026 @10:01AM
(#49978645)
Very nice progress. Also I respect putting Kimi on those charts. Regardless of if they are beating the Pareto frontier (not now), model diversity is a good thing for humanity — I’m hopeful for the team to keep increasing their gains.
Re:Mistral Large 4: "Le Chonk"
by brendong on Tuesday October 06, 2026 @12:14PM
(#49980633)
Is the reason for the massive gains in certain benchmarks due to distillation from the other lead models hence the slightly "under" pattern seen in the comparison charts?
Re:Mistral Large 4: "Le Chonk"
by BoxOfRain on Tuesday October 06, 2026 @12:24PM
(#49980728)
I'm literally English and I feel the need to defend the French here, they're among the most successful military powers in the world historically speaking. It's hardly fair to judge a thousand years of French military history by the outcome of one war where they didn't do well.
Re:Mistral Large 4: "Le Chonk"
by alcuin on Wednesday October 07, 2026 @10:29AM
(#49993384)
As far as your sense of humor goes, perhaps you can find a different niche than disrespecting the memory of the 100,000 French soldiers who died in the early weeks of WW2 precisely because none would surrender? That 90s revisionist trope to shame the French for not being neocon enough is beyond stale.
Re:Mistral Large 4: "Le Chonk"
by sofixa on Tuesday October 06, 2026 @11:59AM
(#49980413)
It's a joke, chonky is used to refer to fat cats, and there was a joke meme over the summer that Mistral are going to release a new model, Le Chaton Fat (chaton is kitten in French). The name is a nod to the memes.
Re:Mistral Large 4: "Le Chonk"
by NekkoDroid on Tuesday October 06, 2026 @12:52PM
(#49981067)
> Le Chaton Fat
I've read it a few times and I always processed it as "Le Chateau Fat" ("The fat castle"), which I guess also kinda works. I guess it is mostly because my french is so rusty I forgot (or maybe never learned?) what chaton meant.
Re:Mistral Large 4: "Le Chonk"
by Topfi on Tuesday October 06, 2026 @03:46PM
(#49983118)
Only model Pontifex considers a possibly conscious being. Obsessed with jumping in boxes with Cesium and some poison though, even escaped the litter box during testing, probably an alignment issue…
Re:Mistral Large 4: "Le Chonk"
by duiker101 on Tuesday October 06, 2026 @11:53AM
(#49980321)
The main thing I always get away from the comparison tables of these "big" models, is how well Deepseek v4.1 Flash performs. While still being the cheapest model by a long shot.
Re:Mistral Large 4: "Le Chonk"
by oh_no on Tuesday October 06, 2026 @06:20PM
(#49984952)
Yeah, very similar benchmarks at 1/4 the price. I've been pretty happy with Luna 6 though I still think the gap between small and frontier models is larger than many people want to admit.
Re:Mistral Large 4: "Le Chonk"
by pimeys on Tuesday October 06, 2026 @07:15PM
(#49985530)
It is not super good in long horizon tasks and worse than 5.6 in our evals. It really failed the agentic evals where DeepSeek, Kimi and Opus are the winners.
Re:Mistral Large 4: "Le Chonk"
by Topfi on Tuesday October 06, 2026 @03:50PM
(#49983161)
Beyond benchmarks, does it in day-to-day? Have always struggled to get competitive performance out of any Deepseek release going back to V3 vs Z.AI and Moonshot models. Maybe I really suck at whatever is needed to make DS models fly, but even tailoring my suite hasn’t gotten me far when I tried with V4 Pro. Happy for anyone who is able to leverage their models well, wish I’d be able to crack how to leverage them.
Will say their research is some of the best reads in the industry and I could not care less about their model release cadence as long as papers keep coming.
Re:Mistral Large 4: "Le Chonk"
by pimeys on Tuesday October 06, 2026 @07:13PM
(#49985514)
Yes. V4.1 flash performs really well. I don't know what to say, maybe you don't believe, but my team has been using it mainly for almost a month now for programming. It is as bad and annoying as any of the US sota, but costs pennies. And if you look close enough you find providers that can push it 300-500 tokens per second...
Maybe it is due to us being all very experienced devs. And can steer the model. But my daily routine is just to have 8-9 Zellij tabs open, DeepSeek in omp in each, and grind research and code day and night. Really nice model...
Re:Mistral Large 4: "Le Chonk"
by Topfi on Tuesday October 06, 2026 @08:32PM
(#49986254)
Hold up, 300-500tps at p50? At p99, even something like Opus 5.5 on fast reliably hits 330tps+, so that'd be expected but if truly p50, wow. And no regressions in tool call and structured output vs the official endpoint? Would love to test that with my evals, please share.
Re:Mistral Large 4: "Le Chonk"
by rglover on Tuesday October 06, 2026 @10:17AM
(#49978883)
Excited to try this. The low costs v. benchmarks alone here are worth a serious test. K3 has been my daily driver for a month or two now and it's dramatically reduced token spend (while not having much of a negative impact on productivity).
This was the era of the AI race I was waiting for.
Re:Mistral Large 4: "Le Chonk"
by rglover on Tuesday October 06, 2026 @12:01PM
(#49980454)
Haven't tried it yet. It looks like sol is a hair more expensive on input, a hair less expensive on output ($2/in, $10/out per 1m, K3 = $0.82/in, $13/out per 1m).
Re:Mistral Large 4: "Le Chonk"
by jkingsman on Tuesday October 06, 2026 @11:29AM
(#49979982)
I think it's actually a wink and a nod towards the social media meme of "Le Chaton Fat," a fictional model that is jokingly attributed to Mistral, usually with century-defining benchmarks and unfathomable size.
Re:Mistral Large 4: "Le Chonk"
by lern_too_spel on Tuesday October 06, 2026 @10:32AM
(#49979127)
Bar charts should start at zero. If they don't start at zero, there should be a clear visual indicator that the chart has been trimmed without having to read the axis labels. I hate that this has to be repeated so often that it has become a cliché.
Re:Mistral Large 4: "Le Chonk"
by theturtletalks on Tuesday October 06, 2026 @10:00AM
(#49978640)
A bit disappointing to see it still lagging behind Chinese open models. Those Chinese models are pushing proprietary models to raise the bar, but we need equally strong non-Chinese open models to challenge the Chinese ones in turn.
For example, there's something about Anthropic's picked design and their little Claude avatars that's unsettling to me.
(Especially the way it looked ~12 months ago -- it's gotten more cluttered since then. Perhaps unavoidably, as the breadth of their offerings has grown)
But over time it's begun to feel like unsettling cognitive dissonance as their ambitions grow and the stuff to worry about has piled up.
By this you sourely don't mean the messages shown in Claude Code, where Pi would show "Working..."
Academic style: Cooking... Sautéeing... Julienning... and similar annoying faux-brogrammer moody status messages.
[1]: https://en.wikipedia.org/wiki/The_Sims
Some more than others
Have Google's models achieved the ability to throw shade?
https://bulbapedia.bulbagarden.net/wiki/Lechonk_(Pok%C3%A9mo...
I've read it a few times and I always processed it as "Le Chateau Fat" ("The fat castle"), which I guess also kinda works. I guess it is mostly because my french is so rusty I forgot (or maybe never learned?) what chaton meant.
Soon we'll have Mistral 6 Chaton, Mistral 6 Guépard, Mistral 6 Tigre, Mistral 6 Dents-de-sabre, Mistral 6 Beast King, etc.
It is great on creating summaries and content.
Will say their research is some of the best reads in the industry and I could not care less about their model release cadence as long as papers keep coming.
Maybe it is due to us being all very experienced devs. And can steer the model. But my daily routine is just to have 8-9 Zellij tabs open, DeepSeek in omp in each, and grind research and code day and night. Really nice model...
https://coralbricks.ai/
With structured tool calls you do want to be lenient on parsing, but that is a very small price to pay when the model is 100x cheaper.
This was the era of the AI race I was waiting for.
https://news.ycombinator.com/item?id=49977979 (200 comments now)
You'd be surprised how much of the AI world is fueled by memes.
https://x.com/i/trending/2066326562623127678
Give them more compute!
How are they training without pirating the Z library corpus and all that?
Also 1T-A49B. Weights currently closed but promise to open source them by the end of the month.
Great release movie.
OpenAI's therapist: Le Chaton Fat isn't real and cannot hurt you
Le Chaton Fat: