Mistral Large 4
Mistral Large 4 (Le Chonk) is Mistral AI's 1-trillion-parameter, natively multimodal, open-weight model, served through the Mistral API and released as downloadable weights.
Our verdict
An open-weight frontier model you can call through Mistral's API or run on your own infrastructure. Mistral calls it the best open-weight model for coding. A strong choice when you want the option to self-host later without switching models.
Best for
- ✓ Agentic coding
- ✓ Self-hosting an open-weight model
- ✓ Defensive security work
Watch out for
- ! Businesses running a Model-as-a-Service offering need a separate commercial license from Mistral
- ! High reasoning is the default; Mistral recommends Low for simple tasks and Medium for everyday ones
Pricing
Read from the official pricing page. Prices in USD; always confirm on the vendor's site.
| Plan | Price | Details |
|---|---|---|
| Mistral API | Pay as you go | $0.70 per 1M input tokens, $0.07 per 1M cached input tokens, $2.10 per 1M output tokens. |
Note: Prices from Mistral's launch post. The weights are free to use, modify, fine-tune and commercialize under the Mistral Ecosystem Open-Weight license, a modified MIT license; self-hosting costs depend on your own hardware.
FAQ
How much does Mistral Large 4 cost?
On the Mistral API it's $0.70 per 1M input tokens, $0.07 per 1M cached input tokens and $2.10 per 1M output tokens. Running the open weights yourself has no license fee for most uses, but you pay for the hardware.
Is Mistral Large 4 open source?
It's open-weight. The weights are released under the Mistral Ecosystem Open-Weight license, a modified MIT license: free to use, modify, fine-tune and commercialize. The one condition is that businesses operating a Model-as-a-Service offering must get a separate commercial license from Mistral first.
Where can I run Mistral Large 4?
Through Mistral Studio, the Mistral API, OpenRouter, Microsoft Foundry and NVIDIA NIM. You can also download the open weights from Hugging Face and serve them yourself with vLLM or SGLang.
What are the reasoning modes?
Three: Low for simple tasks, Medium for everyday tasks, and High, the default, for the hardest tasks that need high precision and thorough reasoning.
What do 1,000 chats cost?
Take 1,000 chats of 1,000 input tokens and 500 output tokens each, with no caching: that's 1M input and 0.5M output tokens. 1 x $0.70 + 0.5 x $2.10 = $1.75 at Mistral API list prices.
Related
Other ai model apis
Get New this week by email
One short email a week with the new releases in AI coding and building tools, each linked to its source. Unsubscribe anytime.
We'll ask you to confirm by email. See our privacy policy.