LLMS • CURATED • UPDATED OCT 7, 2026

Mistral Large 4

Europe's 1-trillion-parameter model, with weights you can download
Score
26.1 · rank 31
Price
Freemium
API
Yes
Weights
Yes

Mistral Large 4, nicknamed Le Chonk, is Mistral's flagship. The public preview opened on October 6, 2026. On October 29 Mistral published the weights of a further-trained checkpoint. The model card describes a natively multimodal mixture-of-experts model with 1.05 trillion total parameters, 52 billion active, a 1.6 billion parameter vision encoder, and a 1 million token context, trained on 3,800 NVIDIA Grace Blackwell GPUs in European datacenters. Mistral's weights post says this checkpoint scores 41% on Terminal-Bench 4.0, up 13 points from the preview, and 48% on the Artificial Analysis Coding Agent Index while using 47% fewer tokens per task than GLM-5.3, which it calls a tie. The same post sets the API price at $0.70 per million input tokens, $0.07 per million cached input tokens, and $2.10 per million output tokens. The October 6 DeepSWE score of 61.7% is the preview checkpoint, which Mistral has retired. Weights are on Hugging Face as mistralai/Mistral-Large-4-1T-A52B.

The weights are out, so a European frontier model can be downloaded and run without Mistral serving it. The licence is a modified MIT: use, changes, and commercial products are allowed, and a business that offers the model as a service needs a separate commercial licence. Coding and cyber figures here are Mistral's, and the DeepSWE number is labelled as the preview.

Pricing Freemium
Platforms web, api, local
API Yes
Open Source Yes
Modalities LLMs, IDEs & Coding Tools, Multimodal Reasoning
Best For Best European Frontier Model
Date Added 2026-10-07
  • ✓Weights published October 29, 2026, on Hugging Face
  • ✓API price of $0.70 input, $0.07 cached input, and $2.10 output per million tokens, per Mistral
  • ✓Three reasoning modes, with High as the default
  • ✓Terminal-Bench 4.0 at 41%, Mistral's figure for the weights release
  • ✓Trained in Europe, and runnable without Mistral serving the API
  • ⚠The DeepSWE 61.7% figure is the October 6 preview, not the weights checkpoint
  • ⚠Other coding and cyber scores on the weights post are Mistral's, not a CuratedAI retest
  • ⚠Model-as-a-service use needs a separate commercial licence from Mistral
  • ⚠A specialized cybersecurity variant is limited to vetted partners and authorized systems
  • ⚠About a trillion parameters. Self-hosting needs a large GPU cluster even with 52 billion active

Call mistral-large-4-0 from Mistral Studio, the API, or Le Chat's Vibe Code. To self-host, download mistralai/Mistral-Large-4-1T-A52B and serve it with vLLM or SGLang. Read the Mistral Ecosystem Open-Weight License before you offer the model as a hosted service. The October 6 preview checkpoint is retired.

Artificial Analysis Intelligence Index v5.0.1 · llm
26.1
Rank #31 · 2026-11-21
DeepSWE v1.1 (Mistral, October 6 preview) · ide-coding
61.7%
Freemium Free tier available

Free tier includes limited features. Paid plans unlock full access, higher usage limits, and commercial usage rights.

View Mistral Large 4 Alternatives (2026) →

Compare Mistral Large 4 with 5+ similar llms AI tools.

❓
Q

Is Mistral Large 4 free?

A

Mistral Large 4 offers a free tier with optional paid upgrades.

Q

Does Mistral Large 4 have an API?

A

Yes, Mistral Large 4 offers an API for programmatic integration.

Q

What is Mistral Large 4 best for?

A

Mistral Large 4 is best for Best European Frontier Model. Mistral Large 4, nicknamed Le Chonk, is Mistral's flagship. The weights are out, so a European frontier model can be downloaded and run without Mistral serving it. The licence is a modified MIT: use, changes, and commercial products are allowed, and a business that offers the model as a service needs a separate commercial licence. Coding and cyber figures here are Mistral's, and the DeepSWE number is labelled as the preview.

Q

What platforms does Mistral Large 4 support?

A

Mistral Large 4 supports web, api, local.

Q

Is Mistral Large 4 open source?

A

Yes, Mistral Large 4 is open source. You can access the source code, contribute, and deploy it on your own infrastructure.

Q

How do I get started with Mistral Large 4?

A

Call mistral-large-4-0 from Mistral Studio, the API, or Le Chat's Vibe Code. To self-host, download mistralai/Mistral-Large-4-1T-A52B and serve it with vLLM or SGLang. Read the Mistral Ecosystem Open-Weight License before you offer the model as a hosted service. The October 6 preview checkpoint is r...

Q

How do I use Mistral Large 4?

A

Mistral Large 4 is a large language model for text generation, analysis, and conversation. Access through the web interface. Enter prompts or questions to get responses. It excels at weights published october 29, 2026, on hugging face.

🏷️

Work on Mistral Large 4? You're hand-reviewed in our directory. Add this badge to your site — it links back to this profile.

Featured on CuratedAI