Skip to content

test: verify MiniMax models through Context7Agent - #2967

Open
octo-patch wants to merge 1 commit into
upstash:masterfrom
octo-patch:octo/20260731-provider-add-recvqf3GrCZnYs
Open

test: verify MiniMax models through Context7Agent#2967
octo-patch wants to merge 1 commit into
upstash:masterfrom
octo-patch:octo/20260731-provider-add-recvqf3GrCZnYs

Conversation

@octo-patch

Copy link
Copy Markdown

Reason: Add a MiniMax OpenAI-compatible fixture that verifies both current models through Context7Agent at the configured regional endpoints.

  • Add @ai-sdk/openai as a development dependency.
  • Add a deterministic fixture test for MiniMax-M3 and MiniMax-M2.7 at the global and China endpoints.
  • Verify that the Context7Agent path sends chat-completion requests with the selected model and Context7 tools.

Checks:

  • pnpm install --frozen-lockfile --ignore-scripts
  • pnpm --filter @upstash/context7-tools-ai-sdk typecheck
  • pnpm --filter @upstash/context7-tools-ai-sdk exec vitest run src/minimax.test.ts
  • ./node_modules/.bin/tsup from packages/tools-ai-sdk
  • ESLint and Prettier checks for the changed files
  • Secret scan passed
  • The full package suite was also run; its five network-backed cases require an unset cloud-region environment variable, while the remaining fifteen tests passed.

@fahreddinozcan

Copy link
Copy Markdown
Collaborator

Hey @octo-patch, what's the motivation of this test? Is there a problem with minimax models on our side?

@octo-patch

Copy link
Copy Markdown
Author

There isn't a known MiniMax-specific problem in Context7. This is deterministic compatibility coverage that verifies Context7Agent accepts the OpenAI-compatible MiniMax provider for both model IDs and regional base URLs, and forwards Context7 tools through the normal agent path without making live API calls.

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.

2 participants