Mistral AI provider implementation.

Uses the OpenAI-compatible API with Mistral-specific extensions:

• reasoning_mode - Enable reasoning mode for complex tasks
• prediction_mode - Enable prediction mode
• safe_prompt - Enable safe prompt filtering

Configuration

Set your API key via environment variable:

export MISTRAL_API_KEY="your-mistral-api-key-here"

Or in config:

config :nous, :mistral,
  api_key: "your-mistral-api-key-here"

Usage

# Via Model.parse
model = Nous.Model.parse("mistral:mistral-large-latest")

# Direct provider usage
{:ok, response} = Nous.Providers.Mistral.chat(%{
  "model" => "mistral-large-latest",
  "messages" => [%{"role" => "user", "content" => "Hello"}]
})

# With reasoning mode
{:ok, response} = Nous.Providers.Mistral.chat(%{
  "model" => "mistral-large-latest",
  "messages" => messages,
  "reasoning_mode" => true
})