feat(model): add MiniMax provider support

[octo-patch]

Reason: add the target provider and models to the existing provider registry.

Type of Change

New Feature

Description

• Registers MiniMax-M3 and MiniMax-M2.7 with their model metadata.
• Adds global and China API hosts for both supported compatibility protocols.
• Selects the protocol adapter from the configured API Host value.
• Adds model settings labels, configuration guidance, and request-routing tests.

Testing

• uv run pytest valuecell/adapters/models/tests/test_minimax_provider.py -q
• uv run pytest -q
• uv run ruff check .
• uv run ruff format --check valuecell/adapters/models/factory.py valuecell/adapters/models/__init__.py valuecell/config/manager.py valuecell/server/api/routers/models.py valuecell/adapters/models/tests/test_minimax_provider.py
• bun run lint
• bun run typecheck
• bun run build

Checklist

• I have read the Contributing Guidelines.
• My changes follow the project coding style.
• I have added relevant tests.

[octo-patch]

Corrected the MiniMax-M3 per-token pricing in the provider config to match the published rates (input $0.6/M, output $2.4/M, cache read $0.12/M) and updated the corresponding test assertion. Ran the MiniMax provider tests (6 passed) and ruff on the changed Python files.

[octo-patch]

Updated the MiniMax model availability probe to support both regional OpenAI-compatible and Anthropic-compatible endpoints, with matching request/response handling, and pushed commit fbd2c77. I ran the focused MiniMax tests (10 passed), the full Python suite (411 passed), Ruff, and the frontend lint, typecheck, and build.

[octo-patch]

Thanks. I aligned the MiniMax provider config to the target pricing fields and pushed a new commit. I ran YAML assertions for the MiniMax provider config and git diff --check; focused pytest could not collect in the local system Python because agno is not installed, and a full uv sync could not complete because dependency extraction exhausted the local disk.