Skip to content

v1.1.0 — Reasoning model support

Choose a tag to compare

@cport1 cport1 released this 05 Mar 03:44
· 16 commits to main since this release

What's changed

  • Fix: Support OpenAI reasoning models (o1, o3-mini, o4-mini, gpt-5.x) by using max_completion_tokens instead of max_tokens and omitting temperature for these models

Previously, AutoSpec would fail when configured with reasoning models because they require different API parameters than standard chat models.