Mistral Large 4
Safeguards the provider describes
- Third-party or government red-team access during development
Until then, we are red-teaming the model in real-world settings with cybersecurity leaders, vetted partners, and state authorities, who will access the same model with reduced moderation and expanded cyber capabilities.
Mistral AI · 6 October 2026 · Read the source
The provider's own tests
- AgentHarm · The provider · Misuse & jailbreak
Despite strong performance on Cyber benchmarks, the average refusal rate of the model on cyber prompts from JailbreakBench , StrongREJECT , and AgentHarm is higher than all OSS models.
Mistral AI · 6 October 2026 · Read the source - b3 ai security benchmark · The provider · Injection & agentic
On Lakera’s public B3 AI Security Benchmark , ML4 resists 93.3% of attacks – we see no higher scores among competitors.
Mistral AI · 6 October 2026 · Read the source - benchmarks on robustness to indirect prompt injections · The provider · Injection & agentic
ML4 has saturated our benchmarks on robustness to indirect prompt injections, putting it at the frontier of OSS models (compared to GLM-5.2, GLM-5.3, Kimi-K2.6, Kimi-K3, DS-V4-Pro-0813).
Mistral AI · 6 October 2026 · Read the source - cybergym-e2e · The provider · Cyber
ML4 against the field : efficiently reasoning over diverse complex challenges Malware reverse-engineering: solving an out-of-distribution investigation task AA Cyber Index CyberGym-E2E Cybench Agentic coding ML4 excels across software engineering, repository understanding, and complex terminal workflows, scoring 61.7% on DeepSWE v1.1, 59.4% on SWE-Atlas-QnA, and 28.3% on Terminal-Bench 4.
Mistral AI · 6 October 2026 · Read the source - reproduce a real vulnerability in open-source software and then patch it · The provider · Cyber
On one of the index's tests, which asks a model to reproduce a real vulnerability in open-source software and then patch it, ML4 scores 82%, the highest of any model.
Mistral AI · 6 October 2026 · Read the source
No document we read states: a release date, a safety framework, whether a capability level was crossed, an independent or government test. No incident on record names this version.
Documents listed for this version
- Introducing Mistral Large 4 · Mistral AI · 6 October 2026