fix(inference): send max_completion_tokens for GPT-5 and o-series models - #6643

Merged

cv merged 10 commits into

NVIDIA:mainfrom

gregmercer:fix/gpt-5.4-max-completion-tokens

Jul 11, 2026

Commits on Jul 10, 2026

fix(inference): send max_completion_tokens for GPT-5 and o-series models

gregmercer
and
claude
committed
4fc9e30

Commits on Jul 11, 2026

fix(inference): cover GPT and o-series model families
cv
committed
b019b59
merge(main): sync GPT family compatibility
cv
committed
ff79965
test(inference): isolate GPT token-field coverage
cv
committed
bc9bc4c
merge(main): refresh GPT-5 token compatibility
cv
committed
91b73e9
fix(inference): omit temperature for reasoning probes
cv
committed
b5eb7ff
Merge branch 'main' into fix/gpt-5.4-max-completion-tokens
cv
authored
ea8132e
Merge branch 'main' into fix/gpt-5.4-max-completion-tokens
cv
authored
d75838d
Merge branch 'main' into fix/gpt-5.4-max-completion-tokens
cv
authored
2fb2b33
Merge branch 'main' into fix/gpt-5.4-max-completion-tokens
cv
authored
0c3fc28