fix: use max_completion_tokens for reasoning models (o1/o3/o4/gpt-5) #544
+39
−5
GitHub Advanced Security / CodeQL
succeeded
Mar 29, 2026 in 2s
No new alerts in code changed by this pull request
Loading