Codestral Configuration File

I have been experimenting with Mistral AI codestral. The coding model that is accessible through API at the following url: codestral.

The model comes with two endpoints: chat and fim(fill-in-the-Middle). I have been able to successfully use these with the following addition to the ~./vibe/config.toml file.

Here are working additions to config.toml:

[[providers]]
name = "codestral"
api_base = "https://codestral.mistral.ai/v1"
api_key_env_var = "CODESTRAL_API_KEY"
api_style = "openai"
backend = "mistral"

[providers.extra_headers]

[[models]]
name = "codestral-latest"
provider = "codestral"
api_base = "https://codestral.mistral.ai/v1/fim/completions"
alias = "codestral-fim"
completion_type = "fim"
temperature = 0.2
thinking = "off"
auto_compact_threshold = 200000
 
[[models]]
name = "codestral-latest"
provider = "codestral"
api_base = "https://codestral.mistral.ai/v1/chat/completions"
alias = "codestral-chat"
completion_type = "chat"
temperature = 0.7
thinking = "off"
auto_compact_threshold = 200000

or append to config.toml:

cat >> config.toml << 'EOF'

[[providers]]
name = "codestral"
api_base = "https://codestral.mistral.ai/v1"
api_key_env_var = "CODESTRAL_API_KEY"
api_style = "openai"
backend = "mistral"

[providers.extra_headers]

[[models]]
name = "codestral-latest"
provider = "codestral"
api_base = "https://codestral.mistral.ai/v1/fim/completions"
alias = "codestral-fim"
completion_type = "fim"
temperature = 0.2
thinking = "off"
auto_compact_threshold = 200000

[[models]]
name = "codestral-latest"
provider = "codestral"
api_base = "https://codestral.mistral.ai/v1/chat/completions"
alias = "codestral-chat"
completion_type = "chat"
temperature = 0.7
thinking = "off"
auto_compact_threshold = 200000
EOF

Alongside the MISTRAL_API_KEY, you'll need to add the CODESTRAL_API_KEY to the ~/.vibe/.env.

A successful addition should result in: