[guillaumegay13]

What this changes

Adds MiniMax-M3.1-Flash for the API key and subscription routes. Each entry carries the generation cap and reasoning_effort (low to max, default max).

Left out on purpose, per the docs: temperature and top_p are ignored on Flash (fixed at 1 and 0.95). reasoning_split: false, thinking.type: disabled and service_tier are rejected or unsupported on Flash.

reasoning_effort only takes effect on Flash, so MiniMax-M3 is unchanged. The subscription entry uses the Chat Completions name like the other MiniMax subscription entries. The gateway maps it to output_config.effort on the Messages API (mnfst/llm-gateway#3090).

Not verified live: the MiniMax key available to me has no balance.

Type of change

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

Source

https://platform.minimax.io/docs/api-reference/text-openai-api.md
https://platform.minimax.io/docs/api-reference/text-anthropic-api.md

Before opening

• npm run validate and npm test pass locally
• Filenames follow the convention: models/<provider>/<model>.yaml, or <model>-subscription.yaml for the subscription route
• No existing parameter was removed (removals are blocked, see CONTRIBUTING)

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	modelparams-dev	6f753c5	Commit Preview URL Branch Preview URL	Oct 29 2026, 10:10 AM