Skip to content

feat: add MiniMax as first-class LLM provider#4038

Open
octo-patch wants to merge 1 commit intoAgenta-AI:mainfrom
octo-patch:feature/add-minimax-provider
Open

feat: add MiniMax as first-class LLM provider#4038
octo-patch wants to merge 1 commit intoAgenta-AI:mainfrom
octo-patch:feature/add-minimax-provider

Conversation

@octo-patch
Copy link

Summary

Add MiniMax AI as a built-in LLM provider alongside existing providers (OpenAI, Anthropic, Google Gemini, etc.) for both standard and custom provider modes.

Changes

Backend (API + SDK):

  • Register MINIMAX in StandardProviderKind and CustomProviderKind enums
  • Add MiniMax models to supported_llm_models via LiteLLM native minimax/ prefix routing:
    • minimax/MiniMax-M2.7 (latest flagship, 1M context)
    • minimax/MiniMax-M2.7-highspeed (optimized for speed)
    • minimax/MiniMax-M2.5 (previous generation)
    • minimax/MiniMax-M2.5-lightning (fast previous generation)

Frontend (Web):

  • Add MINIMAX to SecretDTOProvider enum, PROVIDER_LABELS, and PROVIDER_KINDS
  • Add MINIMAX_API_KEY env var mapping in llmProviders.ts
  • Add MiniMax to llmAvailableProviders list for the settings UI
  • Add MiniMax SVG icon component and register in LLMIconMap
  • Add MiniMax to PROVIDER_ICON_MAP for the provider selector dropdown

Tests:

  • 17 unit tests: model registry, provider mapping, enum validation
  • 10 integration tests: secret parsing roundtrip, provider normalization, enum consistency

How it works

MiniMax provides an OpenAI-compatible API at https://api.minimax.io/v1. LiteLLM already has native MiniMax provider support, so models are routed via the minimax/ prefix. Users configure MiniMax by adding their MINIMAX_API_KEY in the Agenta secrets vault, just like any other standard provider.

Test plan

  • All 27 new tests pass (17 unit + 10 integration)
  • Verify MiniMax appears in provider selector dropdown
  • Verify adding MINIMAX_API_KEY and selecting a MiniMax model works in Playground
  • Verify existing providers are unaffected

Add MiniMax AI as a built-in LLM provider alongside existing providers
(OpenAI, Anthropic, etc.) for both standard and custom provider modes.

Backend:
- Register MiniMax in StandardProviderKind and CustomProviderKind enums
- Add MiniMax-M2.7, M2.7-highspeed, M2.5, M2.5-lightning models to
  supported_llm_models (via LiteLLM's native minimax/ prefix routing)

Frontend:
- Add MiniMax to SecretDTOProvider enum, PROVIDER_LABELS, and PROVIDER_KINDS
- Add MINIMAX_API_KEY env var mapping in llmProviders.ts
- Add MiniMax to llmAvailableProviders list
- Add MiniMax icon component (SVG) and register in LLMIconMap
- Add MiniMax to PROVIDER_ICON_MAP for provider selector dropdown

Tests:
- 17 unit tests for model registry, provider mapping, and enum validation
- 10 integration tests for secret parsing, normalization, and consistency
@vercel
Copy link

vercel bot commented Mar 22, 2026

Someone is attempting to deploy a commit to the agenta projects Team on Vercel.

A member of the Team first needs to authorize it.

@dosubot dosubot bot added the size:L This PR changes 100-499 lines, ignoring generated files. label Mar 22, 2026
@CLAassistant
Copy link

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.


PR Bot seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
You have signed the CLA already but the status is still pending? Let us recheck it.

Copy link
Member

@mmabrouk mmabrouk left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hey @octo-patch

Thank you for your contribution!

  1. Please use the PR template for your PR
  2. Please check the contributing guidelines https://agenta.ai/docs/contributing/first-pr we require for instance a short loom video demo for the feature with the PR

Thank you!

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

Labels

Backend feature Frontend size:L This PR changes 100-499 lines, ignoring generated files. tests

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants