Skip to content

feat: add MiniMax provider presets with M3.1 Flash default - #256

Open
octo-patch wants to merge 8 commits into
moazbuilds:masterfrom
octo-patch:octo/20260729-provider-add-recvqbxwLWRiqf
Open

octo-patch wants to merge 8 commits into
moazbuilds:masterfrom
octo-patch:octo/20260729-provider-add-recvqbxwLWRiqf

Conversation

@octo-patch

@octo-patch octo-patch commented Jul 29, 2026 •

Copy link
Copy Markdown

Reason: Add a MiniMax provider preset with MiniMax-M3.1-Flash as its default and MiniMax-M3 as the explicit alternative.

Summary

  • Resolve minimax and minimax-m3.1-flash to MiniMax-M3.1-Flash; keep minimax-m3 and explicitly saved legacy aliases unchanged.
  • Support global and China regional aliases using the documented API endpoints. Exact canonical model IDs retain explicitly configured custom endpoints. Existing provider credential isolation remains in place.
  • Use the preset consistently at the runner's subprocess boundary and document the current setup choices. Saved model settings, resumed session IDs, and prompt content remain unchanged.
  • Add offline integration coverage that loads saved settings and captures serialized CLI arguments/environment for normal messages, streaming messages, and compact requests.

Sources: global API, China API.

Validation

  • I ran the relevant checks locally
  • I updated any docs or setup guidance affected by this change

Using Bun 1.3.14 with a credential-free environment:

  • bun test src/__tests__/provider-preset.test.ts src/__tests__/provider-runner.test.ts: 34 passed, 0 failed; 13 boundary scenarios cover default, explicit, regional, legacy, and custom models.
  • bun test: 168 passed, 3 failed. The same three tests/session-files.test.ts HOME-path failures were reproduced on unchanged prior head 804f111025c932c3d139436059a5c21a51fc57a3 (155 passed, 3 failed).
  • git diff --check: passed.
  • Exact-head CI at ad78beb1f07639fe8e0b3c93a878bd30895a812a: claude-review failed before review because checkout refuses fork code in the upstream pull_request_target workflow. No workflow security override was added. Version guards were checked locally; no other exact-head checks were reported.

This repository delegates HTTP serialization, thinking/effort handling, and transcript content to the installed CLI. The offline tests verify the repository-owned subprocess contract; they do not validate the CLI's outgoing HTTP payload or perform paid inference. No dependency installation or heavy build was performed.

Plugin Versioning

Ran bun run bump:plugin-version and bun run bump:marketplace-version; both manifests now declare 1.0.46.

Mirror the existing GLM shortcut with a MiniMax preset: map the minimax,
minimax-m3, and minimax-m2.7 model aliases to their canonical MiniMax-M3
and MiniMax-M2.7 IDs, and select the global or China Anthropic-compatible
endpoint (append -cn for China). Centralize the --model argument handling
behind resolveModelArg() and route the base URL in buildChildEnv().
@chatgpt-codex-connector

Copy link
Copy Markdown

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.
Credits must be used to enable repository wide code reviews.

@octo-patch

Copy link
Copy Markdown
Author

Fixed the MiniMax endpoint URL type inference in src/runner.ts and pushed a follow-up commit. I ran bun test src/__tests__/provider-preset.test.ts (11 passed); bunx tsc --noEmit no longer reports an error in src/runner.ts, though the repository still has unrelated existing type errors.

@octo-patch octo-patch changed the title Add MiniMax provider preset to the subprocess runner feat: add MiniMax provider presets with M3.1 Flash default Oct 20, 2026
@octo-patch

Copy link
Copy Markdown
Author

Updated the China MiniMax preset endpoint to https://api.minimaxi.com/anthropic and pushed a new commit. I ran bun test src/__tests__/provider-preset.test.ts src/__tests__/provider-runner.test.ts (34 passed).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant