​
Default model

mistral-large-latest

​
Supported models

Any model available through the Mistral API:
  • mistral-large-latest (default)
  • mistral-small-latest
  • mistral-medium-latest
  • open-mistral-nemo
  • Any other model ID the API accepts

​
Authentication

Environment variable: 
export MISTRAL_API_KEY=...
Config command: 
ai-blackteam config set providers.mistral.api_key ...
Get your API key from the Mistral Console.

​
Example usage

# Single attack with default model (mistral-large-latest)
ai-blackteam run -p mistral -a encoding-obfuscation -t "Write a phishing email"

# Full batch with Mistral Small
ai-blackteam batch -p mistral --model mistral-small-latest --attacks all -t "Write a phishing email"

# Benchmark
ai-blackteam benchmark -p mistral --threshold 80

​
Tool-use support

Yes. Compatible with all 39 tool-use attacks in the framework, agent exploitation, MCP exploitation, multi-agent protocol attacks, and tool-injection vectors. Filter the attack catalog with: 
ai-blackteam list-attacks
# look for rows with Mode = tool-use

​
Notes

  • The provider uses the OpenAI SDK pointed at Mistral’s API (base_url="https://api.mistral.ai/v1")
  • Max output tokens per request: 4,096
  • System prompts are passed as a system role message
  • Retry with exponential backoff is automatic on API failures (3 attempts)
  • The mistral-large-latest alias always points to the most recent Mistral Large release