Skip to content

fix: resolve bare model names to connection defaultModel before upstream calls#474

Open
anuragg-saxenaa wants to merge 4 commits intodecolua:masterfrom
anuragg-saxenaa:master
Open

fix: resolve bare model names to connection defaultModel before upstream calls#474
anuragg-saxenaa wants to merge 4 commits intodecolua:masterfrom
anuragg-saxenaa:master

Conversation

@anuragg-saxenaa
Copy link
Copy Markdown
Contributor

Summary

  • getProviderCredentials now returns defaultModel from the selected connection
  • handleChatCore and handleEmbeddingsCore use the resolved real model ID instead of bare alias names (e.g. "auto") when no "/" separator is present
  • Prevents API rejections when MiniMax connections use bare model names in their aliases

Test plan

  • Send a chat request using a bare model name (e.g. "auto") through a MiniMax connection — verify the upstream receives the real model ID
  • Send a chat request using an explicitly qualified model (e.g. "minimax/auto") — verify behavior is unchanged
  • Test embeddings endpoint with bare model names

🤖 Generated with Claude Code

anuragg-saxenaa and others added 4 commits April 1, 2026 23:33
…eam calls

When a connection uses a bare model name (e.g. "auto", "pro") in its alias,
the qualified model string like "minimax/auto" was sent upstream — but "auto"
is not a valid MiniMax model ID. This caused API rejections.

Now getProviderCredentials returns defaultModel from the selected connection,
and both handleChatCore/handleEmbeddingsCore receive the resolved real model ID
instead of the alias. Only applies to bare names (no "/" separator).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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