v1.104.4

BerriAI/litellmv1.104.4Oct 20, 2026by litellm-dispatcher[bot]

Upgrading?

AI Summary

This release focuses on stability and bug fixes for the LiteLLM proxy and token counting features. It includes the verification of Docker image signatures using Cosign and backports several fixes to the stable/1.104.x branch to ensure compatibility and correct behavior.

Key Highlights

  • Docker images are now cryptographically verified using Cosign for enhanced security.
  • Fixed Guardrail settings loading by restoring `supported_decisions_providers`.
  • Backported team-scoped batch model access support to the stable branch.
  • Resolved Google request compatibility issues in token counting.
  • Restored legacy token count input precedence.

Full Release Notes

## Verify Docker Image Signature

All LiteLLM Docker images are signed with [cosign](https://docs.sigstore.dev/cosign/overview/). Every release is signed with the same key introduced in [commit `0112e53`](https://github.com/BerriAI/litellm/commit/0112e53046018d726492c814b3644b7d376029d0).

**Verify using the pinned commit hash (recommended):**

A commit hash is cryptographically immutable, so this is the strongest way to ensure you are using the original signing key:

```bash
cosign verify \
  --key https://raw.githubusercontent.com/BerriAI/litellm/0112e53046018d726492c814b3644b7d376029d0/cosign.pub \
  ghcr.io/berriai/litellm:v1.104.4
```

**Verify using the release tag (convenience):**

Tags are protected in this repository and resolve to the same key. This option is easier to read but relies on tag protection rules:

```bash
cosign verify \
  --key https://raw.githubusercontent.com/BerriAI/litellm/v1.104.4/cosign.pub \
  ghcr.io/berriai/litellm:v1.104.4
```

Expected output:

```
The following checks were performed on each of these signatures:
  - The cosign claims were validated
  - The signatures were verified against the specified public key
```

---
## What's Changed
* fix(guardrails): restore supported_decisions_providers on stable/1.104.x so Add Guardrail settings load by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45935
* fix(proxy): backport #47067 team-scoped batch model access to stable/1.104.x for v1.104.4 by @mateo-berri in https://github.com/BerriAI/litellm/pull/47344
* fix(count_tokens): backport October 19 fixes to stable/1.104.x by @mateo-berri in https://github.com/BerriAI/litellm/pull/47389
* fix(count_tokens): restore Google request compatibility on stable/1.104.x by @mateo-berri in https://github.com/BerriAI/litellm/pull/47446
* fix(proxy): restore legacy token count input precedence by @joshua-berri in https://github.com/BerriAI/litellm/pull/47454


**Full Changelog**: https://github.com/BerriAI/litellm/compare/v1.104.3...v1.104.4

Related Repositories