[Feat] Add MiniMax-M3 model

[zoomote]

Opened on behalf of Naved Merchant. Follow up by mentioning @zoomote, in the web UI, or in Discord.

Related GitHub Issue

Closes: #888

Description

Adds MiniMax-M3 to the MiniMax provider and makes it the default model, following the same pattern used for the M2.5 and M2.7 additions.

Key implementation details:

• New MiniMax-M3 entry in packages/types/src/providers/minimax.ts: 1M token context window, 16,384 max output tokens, native multimodal input (images, plus video via MiniMax's Anthropic-compatible API), prompt caching enabled, and the same tool configuration as the other MiniMax models (search_and_replace included, apply_diff excluded, reasoning preserved).
• Pricing reflects MiniMax's published pay-as-you-go standard tier with the permanent 50% discount: $0.30/M input, $1.20/M output, $0.06/M cache reads. Input, output, and cache read prices double for requests over 512k input tokens; this is called out in the model description because the tiered-pricing metadata field (longContextPricing) is only applied by the OpenAI cost path, while the MiniMax handler calculates cost via the Anthropic path.
• MiniMax does not publish an M3 cache-write price, so cacheWritesPrice follows the existing M2.x convention ($0.375/M).
• minimaxDefaultModelId updated from MiniMax-M2.7 to MiniMax-M3, matching the established convention of defaulting to the latest MiniMax model.
• Specs grounded in MiniMax's official documentation (models intro, Anthropic API reference, and pay-as-you-go pricing pages).

Reviewers may want to spot-check the pricing values against https://platform.minimax.io/docs/guides/pricing-paygo.

Test Procedure

• Added MiniMax-M3 coverage to src/api/providers/__tests__/minimax.spec.ts: a model-selection test asserting the M3 configuration and a Model Configuration property test; updated the default-model test to expect M3.
• pnpm test from the repo root passes: 414 test files, 6,785 tests green.
• The webview useSelectedModel tests resolve the MiniMax default model dynamically, so they pass unchanged (32 tests).
• Reviewers can verify by selecting the MiniMax provider in settings and confirming MiniMax-M3 appears in the model dropdown as the default.

Pre-Submission Checklist

• Issue Linked: This PR is linked to an approved GitHub Issue (see "Related GitHub Issue" above).
• Scope: My changes are focused on the linked issue (one major feature/fix per PR).
• Self-Review: I have performed a thorough self-review of my code.
• Testing: New and/or updated tests have been added to cover my changes (if applicable).
• Documentation Impact: I have considered if my changes require documentation updates (see "Documentation Updates" section below).
• Contribution Guidelines: I have read and agree to the Contributor Guidelines.

Screenshots / Videos

No UI changes — this is a model registry entry plus provider tests. Zoo Code is a VS Code extension; the new model surfaces in the MiniMax provider's model dropdown in settings.

Documentation Updates

• No documentation updates are required. The MiniMax provider docs page does not enumerate individual models.

Additional Notes

MiniMax has not published an M3 cache-write price; the $0.375/M value follows the existing M2.x convention and can be revisited once official M3 cache-write pricing is available.

Get in Touch

Opened via Roomote on behalf of the requester — follow up via the linked task thread in the provenance note above.

[zoomote]

No code issues found. See task

_{Reviewed 148b428}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!