Skip to content

feat(provider): add deepseek-v4 per-family adapter - #1374

Merged
TheGreatAxios merged 4 commits into
mainfrom
cl-10243-deepseek-v4-per-family-adapter
Oct 9, 2026
Merged

TheGreatAxios merged 4 commits into
mainfrom
cl-10243-deepseek-v4-per-family-adapter

Conversation

@TheGreatAxios

@TheGreatAxios TheGreatAxios commented Oct 9, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

Adds a deepseek-v4 per-family adapter (bifrost-style) registered in the Interchange registry and selected by /^deepseek-v4/i. Removes the V4 branch from the generic openai-compatible adapter and folds deepseek-v4 into detectModelFamily so callers can branch on the family.

Verification

  • 15/15 adapter unit tests pass.
  • Non-V4 models take the untouched generic path (byte-identical, verified by openai-compatible-adapter tests).
  • A V4 stream reports reasoning_tokens usage.

Fixes CL-10241

Lift the DeepSeek V4 wire quirks out of the generic openai-compatible
adapter into a bifrost-style deepseek-v4 adapter, registered in the
AdapterManifest and selected by a deepseek-v4 provider id stamped on
the source for /^deepseek-v4/i models. Removes the V4 branch from the
generic adapter, keeping the pre-V4 reasoning_content strip and the NIM
null-delta patch. Relocates the V4 effort/settings tests.
@linear-code

linear-code Bot commented Oct 9, 2026 •

Copy link
Copy Markdown

CL-10243

CL-10241

@TheGreatAxios
TheGreatAxios marked this pull request as ready for review October 9, 2026 06:05
@TheGreatAxios
TheGreatAxios merged commit 35b2346 into main Oct 9, 2026
13 checks 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