Model Overview & Capabilities
Mistral Large 4 is Mistral AI’s largest frontier Mixture-of-Experts architecture released in public preview on October 6, 2026, claiming Rank #8 on the Vibe Coder Journal Leaderboard with a composite index score of 45.6. Built with approximately 1 trillion total parameters and 52 billion active parameters per token, it delivers verified scores of 46.5% on SWE-bench Verified, 45.0% on Terminal-Bench 2.0, 46.0% on Aider Polyglot, and 87.0% on GPQA Diamond.
Mistral lists standard commercial pricing at $1.36 per million input tokens and $4.18 per million output tokens, which resolves to an effective blended price of $2.065 per million tokens under 3:1 weighting, with promotional launch discounts reducing costs further during preview. At standard pricing, Mistral Large 4 yields an outstanding value score of $0.0453 per index point, significantly undercutting American proprietary frontier models while providing equivalent reasoning depth.
With a 1,000,000-token context window, 65,536 maximum output tokens, and throughput of 92 tokens per second, Mistral Large 4 offers exceptional multilingual coding capabilities, European regulatory data residency compliance, and dependable full-stack backend refactoring for international enterprise environments.
