Today we’re announcing a new collaboration with @nvidia to bring AI to industrial engineering:

- faster simulation
- open benchmarks and supporting datasets for physics AI
- evaluations for agentic AI in CAE and EDA

More here: mistral.ai/news/mistral-n…

Mistral has released new Speech models! Voxtral Small Transcribe 3 takes the #3 spot on the Artificial Analysis Speech to Text leaderboard, and Voxtral 3 TTS enters the top 12 in both Text to Speech arenas

@MistralAI has released Voxtral 3: a new Text to Speech model and three Small Transcribe 3, Mini Transcribe 3 and Mini Transcribe Realtime 3). We evaluated all of them.

Key takeaways:

➤ Speech to Text: Voxtral Small Transcribe 3 takes the #3 spot on the Artificial Analysis Speech to Text leaderboard with a 2.1% AA-WER Index, ahead of ElevenLabs Scribe v2 (2.2%) and Grok Voice Transcribe 2.0 (2.3%). Voxtral Mini Transcribe 3 lands at #5 with 2.2%.

➤ Streaming Speech to Text: Voxtral Mini Transcribe Realtime 3 scores 3.3% final-transcript AA-WER at the 960 ms setting (#6 of 46) and 3.6% at 480 ms, down from 5.2% for the previous Voxtral Mini Transcribe Realtime.

➤ Text to Speech, Provider Voice: Voxtral 3 TTS scores an Elo of 1,218 across 1,599 arena appearances, ranking #12 of 98, just behind Breeze TTS 2 at 1,221 and ahead of Inworld Realtime TTS-2 Flash and StepAudio 2.5 TTS at 1,216. It ranks #13 in English (US) and #12 in English (UK).

➤ Text to Speech, Controlled Voice: Voxtral 3 TTS scores 1,068 across 1,614 appearances, ranking #12 of 43 with the same 8 cloned voices, behind Eleven v3 at 1,072 and level with Ant Group's Ming-3.0-tts at 1,068. It ranks #10 in Assistants and #10 in English (UK).

➤ Pronunciation Robustness: Voxtral 3 TTS scores 84.4%, ranking #13 of 32 models, behind Qwen-Audio-3.1-TTS-Plus at 84.6% and ahead of HeyGen Voice at 83.2%. It ranks #3 in Standalone Terms at 93.9%.

➤ Pricing: Voxtral 3 TTS is priced at $13/1M characters, compared to Gemini 3.8 Flash TTS at $16.5/1M characters, Eleven v4 Turbo at $40/1M characters and Simba 3.2 at $6.6/1M characters.

See more details and listen to samples below ⬇️

🎙️ Introducing Voxtral 3 !

A new family of SOTA speech models:
ASR: Mini & Small for batch transcription, plus real-time streaming
TTS: expressive, multilingual speech, batch + streaming
Voice agent SDK: low-latency conversations + tool calling

Available via API or on your own infrastructure !

We've just released a comprehensive suite of state-of-the-art voice models mistral.ai/news/voxtral-3/. Voice will be the most important interface to AI systems, and we deploy voice models physically close to reasoning models

Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵

FULL INTERVIEW: CEO of Mistral @arthurmensch on building or surrendering, European AI independence, funding gigawatt scale infrastructure in Europe, Le Chonk, AGI, the singularity, and existential risk.

00:24 - Build or Surrender
01:13 - Value in AI application, model and infra s full-stack strategy
05:56 - Open-source & European AI independence
08:31 - Funding gigawatt-scale infrastructure in Europe
11:40 - Overcoming pessimism & sovereign AI
13:26 - AGI, the Singularity, and existential risk concerns
14:50 - AI risks: agent security, control, and misuse
16:16 - The story behind the model name "Le Chonk"

It’s out btw 👀

The duality of a man: one is work, the other is personal.
I'm sad that i don't have access to my dev-box for a couple of weeks (that's where I cook the best stuff) so results are incomplete :'c

Oh no, now i have maxxed out my laptop as well.
I desperately need more compute :'c
(or to touch some grass)

I have maxxed out my dev server... what now ?

Getting there, my beloved DGX spark stay strong it's almost over (it's just the beginning)

Look up! Mistral models are in orbit 🌍.

Proud to join @orbitworksspace and @LoftOrbital's first-of-its-kind AI satellite constellation with @NVIDIAAIInfra.

🇫🇷 Le Chonk joins the sudo L7 leaderboard 🇫🇷

We just added @MistralAI's Mistral Large 4 to sudo L7, our benchmark for staff-level software engineering.

Mistral Large 4 scores 13.3%, ahead of Muse Spark 1.3 (12.2%) and Grok 4.7 (10.6%).

sudo L7 tests whether coding agents can ct code to exercise the judgment of experienced engineers: navigating messy systems, spotting hidden risks, and knowing when technically correct code is still the wrong thing to ship.

Allez, le Chonk ! 🥖

Leaderboard ↓
surgehq.ai/benchmarks/sud…

Solution to my skill issue ?
Throw auto-research at it and learn from it :D

Aight my recipe is ass, it works really well in one direction but is cooked in the other.

Fine tuning is hard :'c

wdym rust compiler ??? what is bro cooking 🤦‍♂️

I literally just stopped a fine tune to do a baseline eval and see if i have to swap the base model 💀

You guys cook too quickly, I (my hardware and wallet) can't keep up

@vllm_project @huggingface yo @NVIDIAAI help a brother out

Prompted Le Chonk to re-make minecraft, not bad at all. Can't wait for the final checkpoint 🐈

I just built a custom harness running entirely on @Cloudflare infrastructure!

- Sessions stored in DO
- Media stored in R2
- Every model through AI gateway
- Dynamic workers for code mode

Check it out below 👇