Skip to content

v3.1.1

Choose a tag to compare

View all tags
@Oaklight Oaklight released this 12 Jun 22:01
· 101 commits to master since this release
v3.1.1
db2a3b0

v3.1.1 — 2026-06-12

Added

  • Periodic model list auto-refresh: The model registry now automatically refreshes from upstream on a configurable interval (default: every 24 hours). Configure via model_refresh_interval_hours in config.yaml (set to 0 to disable). Refresh events are logged at INFO level
  • Fallback model warning: When a model name cannot be resolved, a WARNING log is now emitted showing the requested name and the fallback model used

Changed

  • Default fallback model: Changed from argo:gpt-4o to argo:gpt-5-nano for unresolved chat model names
  • Updated default model list to match the live /v1/models endpoint:
    • Added: GPT-5 family (gpt-5, gpt-5-mini, gpt-5-nano, gpt-5.1, gpt-5.2, gpt-5.4, gpt-5.5), Gemini 3.x (gemini-3.5-flash, gemini-3.1-flash-lite), Claude opus 4.1/4.5/4.6, sonnet 4.5/4.6, haiku 4.5
    • Removed: GPT-3.5 family, GPT-4/4-turbo, gpt-4o-latest, o1-mini, o1-preview, old Claude versions
  • Tiktoken encoding mapping: Added gpt5 and gpt41 prefix mappings, removed gpt3
  • Cleared NO_SYS_MSG_PATTERNS: o1-mini and o1-preview (which couldn't handle system prompts) are retired

Full Changelog: v3.1.0...v3.1.1

Assets 3
Loading