fix(vlm): add max_tokens parameter to VLM completion calls to prevent vLLM rejection#689

Merged

qin-ctx merged 2 commits intovolcengine:mainfrom

mvanhorn:osc/674-vlm-max-tokens

Mar 18, 2026

Commits on Mar 17, 2026

fix(vlm): add max_tokens parameter to VLM completion calls to prevent vLLM rejection

mvanhorn
and
claude
committed
fix(vlm): default max_tokens to None to preserve provider behavior

mvanhorn
and
claude
committed