feat: add MiniMax as inference provider with M3 as default#272
Open
octo-patch wants to merge 4 commits into
Open
feat: add MiniMax as inference provider with M3 as default#272octo-patch wants to merge 4 commits into
octo-patch wants to merge 4 commits into
Conversation
Add MiniMax (api.minimax.io) as a first-class inference provider alongside
OpenAI, Anthropic, and others. Both MiniMax-M2.5 (reasoning tier) and
MiniMax-M2.5-highspeed (fast tier) are now available for model selection.
Changes:
- Add "minimax" to ModelProvider type and minimaxApiKey to config
- Register MiniMax models in provider-registry defaults (enabled)
- Add MiniMax backend routing in inference client with proper
temperature clamping (MiniMax rejects temperature=0)
- Seed MiniMax models in STATIC_MODEL_BASELINE with correct
parameterStyle ("max_tokens")
- Wire MINIMAX_API_KEY through setup wizard, configure menu,
agent loop env export, and CLI entrypoint
- Add provider-registry tests for MiniMax provider
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The STATIC_MODEL_BASELINE now includes MiniMax models, so the valid provider list must include "minimax".
- Add MiniMax-M2.7 and MiniMax-M2.7-highspeed to model list - Set MiniMax-M2.7 as default model - Keep all previous models (M2.5, M2.5-highspeed) as alternatives - Update related unit tests
octo-patch
changed the title
octo-patch
force-pushed
the
feat/add-minimax-provider
branch
from
627776e to
242e7c8
Compare
- Add MiniMax-M3 to model list and set as default (524K context, 128K output, vision support) - Keep MiniMax-M2.7 and MiniMax-M2.7-highspeed as alternatives - Remove older models (MiniMax-M2.5 / MiniMax-M2.5-highspeed) - Update related unit tests and static model baseline - Update README mention from M2.7 to M3
octo-patch
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds MiniMax as a first-class inference provider and upgrades the default model to MiniMax-M3.
Changes
minimaxprovider insrc/inference/provider-registry.tsusing the OpenAI-compatible endpoint athttps://api.minimax.io/v1andMINIMAX_API_KEY.STATIC_MODEL_BASELINE(src/inference/types.ts).minimaxApiKeyplumbing throughsrc/conway/inference.ts,src/config.ts,src/types.ts,src/agent/loop.ts,src/setup/wizard.ts, andsrc/setup/configure.ts(CLI prompts and config flow).src/__tests__/inference/provider-registry.test.tsto assert M3 as the first/default model.Why
M3 is MiniMax's latest generation: 512K context window, 128K max output, and image input support. It supersedes M2.7/M2.5 and is the right default for Conway Automaton's autonomous-agent workload.
Testing
vitest run src/__tests__/inference/provider-registry.test.ts— 34/34 tests passing.MiniMax-M3for thereasoningtier by default.minimaxis included in fallback candidates for bothreasoningandfasttiers.