Mistral Large 4 enters public preview; API is live, open weights due end of October

01 What happened

Mistral AI has launched a public preview of Mistral Large 4 (ML4). The company describes it as a natively multimodal model with 1 trillion parameters, 52 billion of them active, and as its largest and most capable model to date. The preview API is available now on Mistral Studio, and Mistral says model weights will be released by the end of October 2026.

02 Key details

• Mistral AI lists pricing at $1.36 per million input tokens and $4.18 per million output tokens. Access today is through the preview API only, and self-hosting will require the weight release.
• Mistral AI's own reported results include 82% on the reproduce-and-patch test of the Artificial Analysis Cyber Index, 93% of the 40 Cybench challenges and 61.7% on DeepSWE v1.1. The comparisons and conclusions are Mistral's own.
• Until the weights are released, Mistral is red-teaming the model in real-world settings with cybersecurity leaders, vetted partners and state authorities. These groups will get access to the same model with reduced moderation and expanded cyber capabilities.

03 Why it matters

The release brings to market a large model that Mistral plans to publish with open weights and positions as strong in cybersecurity, coding and agentic tasks. The API can be tested now; the figures are Mistral's own, although some results come from independent indexes and third-party evaluators.

04 Who it matters to

Software developers, information security specialists and enterprise AI customers.