Change8

1.84.0-dev.1

📦 litellmView on GitHub →
4 features🐛 13 fixes🔧 14 symbols

Summary

This release introduces Docker image signing verification via cosign and includes numerous fixes across caching, logging, pricing, and provider integrations. A feature to lazy-load optional routers was introduced and subsequently reverted.

Migration Steps

  1. Users running LiteLLM Docker images should verify image signatures using the provided cosign commands to ensure image authenticity.

✨ New Features

  • Added --timeout_worker_healthcheck flag for uvicorn worker triage in the proxy.
  • Added cleanup job for expired LiteLLM dashboard session keys.
  • Added retry settings for generic API logger.
  • Lazy-load optional feature routers on first request (Note: This feature was reverted in a subsequent commit, see bug_fixes).

🐛 Bug Fixes

  • Cached GCP IAM token in Redis to prevent async event loop blocking.
  • Fixed NOMA v2 deepcopy crashing in build scan payload.
  • Used stored-credentials endpoint for tools fetch on MCP edit page.
  • Enabled support for CircleCI rerun failed tests for local_testing jobs.
  • Updated GPT-5.5 Pro Pricing.
  • Preserved prompt_tokens_details through embedding cache round-trip.
  • Backfilled streaming hidden response cost.
  • Reused Anthropic messages config instances for Vertex AI.
  • Preserved items on array branches in anyOf with null in Vertex AI and de-flaked related tests.
  • Replaced deprecated Bedrock Claude 3.7 Sonnet model ID in tests.
  • Cached LiteLLM_Config param reads in DualCache and batch operations.
  • Unified cost calculation in success_handler dict and typed branches.
  • Reverted the feature to lazy-load optional feature routers on first request.

Affected Symbols