Add Xiaomi MiMo V2.5 parameters

[guillaumegay13]

What this changes

Adds configurable request parameters for Xiaomi MiMo V2.5 (api_key).

Covers the native MiMo Chat Completions parameters: max_completion_tokens, thinking.type (thinking is on by default; set disabled to answer directly), temperature, top_p, and response_format.type. Per the deep-thinking docs, temperature/top_p are forced to 1.0/0.95 while thinking is enabled, so they carry an except: { thinking.type: enabled } applicability rule.

Type of change

• Add a model (new YAML file under models/)

Source

• Deep thinking: https://mimo.mi.com/docs/en-US/quick-start/usage-guide/text-generation/deep-thinking
• Rate limit / API: https://mimo.mi.com/docs/en-US/api/guidance/rate-limit
• Cross-checked against OpenRouter supported_parameters for xiaomi/mimo-v2.5.

Smoke-tested live through a local Manifest server routing to the real provider (OpenRouter xiaomi/mimo-v2.5): max_tokens/temperature/top_p/response_format and reasoning all accepted (200).

Closes #71

[vercel]

@guillaumegay13 is attempting to deploy a commit to the Manifest Team on Vercel.

A member of the Team first needs to authorize it.