v1.83.9-nightly

## 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:v1.83.9-nightly
```

**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/v1.83.9-nightly/cosign.pub \
  ghcr.io/berriai/litellm:v1.83.9-nightly
```

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: harden CORS credentials, create_views exception handling, and spend log cleanup loop by @shreyescodes in https://github.com/BerriAI/litellm/pull/25559
* fix(azure/passthrough): populate standard_logging_object via logging hook by @michelligabriele in https://github.com/BerriAI/litellm/pull/25679
* feat(mcp): expose per-server InitializeResult.instructions from gateway by @milan-berri in https://github.com/BerriAI/litellm/pull/25694
* fix(caching): add Responses API params to cache key allow-list by @michelligabriele in https://github.com/BerriAI/litellm/pull/25673
* feat(health-check): add BACKGROUND_HEALTH_CHECK_MAX_TOKENS env var by @Sameerlite in https://github.com/BerriAI/litellm/pull/25344
* feat(anthropic): retry /v1/messages after invalid thinking signature by @Sameerlite in https://github.com/BerriAI/litellm/pull/25674
* fix(ui): delete policy attachments via controlled modal by @Lucas-Song-Dev in https://github.com/BerriAI/litellm/pull/25324
* fix(proxy): enforce team membership in team-scoped key management checks by @milan-berri in https://github.com/BerriAI/litellm/pull/25686
* [Infra] Guard main to only accept PRs from staging and hotfix branches by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25733
* [Infra] Merge main into litellm_internal_staging by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25747
* fix(bedrock/anthropic): accurate cache token cost breakdown in UI and SpendLogs by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25735
* build(ui): rebuild dashboard for cache token cost breakdown changes by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25801
* refactor(ui): reduce Tremor usage in Guardrails Monitor layout by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/25803
* bedrock api response null type handling by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25810
* bedrock api response null type handling by @kothamah in https://github.com/BerriAI/litellm/pull/24147
* [Infra] Expand CI branch filters for non-main PR targets by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25819
* [Refactor] UI - Deleted Keys/Teams: Enterprise notice banner by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25814
* security: prevent API key leaks in error tracebacks, logs, and alerts by @stuxf in https://github.com/BerriAI/litellm/pull/25117
* fix(guardrails): replace custom_code sandbox with RestrictedPython by @jaydns in https://github.com/BerriAI/litellm/pull/25818
* [Infra] Remove unused publish_proxy_extras and prisma_schema_sync jobs from CircleCI config by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25821
* [Infra] Merge dev branch by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25796
* fix: make PodLockManager.release_lock atomic compare-and-delete (re-land #21226) by @joereyna in https://github.com/BerriAI/litellm/pull/24466
* Optimize database query to prevent OOM errors during health checks by @harish876 in https://github.com/BerriAI/litellm/pull/25732
* fix(logging): preserve provider response headers in StandardLoggingPayload by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25807
* Litellm ishaan april15 by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25813
* Litellm ishaan april15 2 by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25828
* docs(troubleshoot): cost discrepancy debugging guide by @Sameerlite in https://github.com/BerriAI/litellm/pull/25622
* Litellm ishaan april14 by @ishaan-berri in https://github.com/BerriAI/litellm/pull/25699
* fix(proxy): harden request parameter handling by @stuxf in https://github.com/BerriAI/litellm/pull/25827
* fix(proxy): add shared path utilities, prevent directory traversal by @stuxf in https://github.com/BerriAI/litellm/pull/25834
* fix(proxy): fix virtual key projected-spend soft budget alerts by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/25838
* fix(ui): render guardrail optional_params bool defaults in Select by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/25806
* fix(ui): use antd Select for MCP ToolTestPanel bool inputs by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/25809
* fix(ui): invalidate org queries after team mutations by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/25812
* litellm oss staging 04/15/2026 by @krrish-berri-2 in https://github.com/BerriAI/litellm/pull/25831
* Litellm day 0 opus 4.7 support by @Sameerlite in https://github.com/BerriAI/litellm/pull/25867
* [Infra] Merge dev branch by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25871
* Fix version in docs by @Sameerlite in https://github.com/BerriAI/litellm/pull/25875
* [Infra] Bump llm_translation_testing resource class to xlarge by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25887
* Litellm hotfix opus 4.7 by @Sameerlite in https://github.com/BerriAI/litellm/pull/25876
* [Infra] Reduce llm_translation_testing parallelism and tolerate worker restarts by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25898
* Add docs announcement bar for Trivy compromise resolution by @krrish-berri-2 in https://github.com/BerriAI/litellm/pull/25870
* [Test] Remove dead Bedrock clear_thinking interleaved-thinking-beta assertion by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25913
* [Test] Mock Bedrock Moonshot tests + [Fix] TogetherAIConfig recursion by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25920
* bump: proxy extras version 0.4.65 → 0.4.66 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25873
* bump: version 1.83.8 → 1.83.9 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25872
* [Infra] Promote Internal Staging to main by @yuneng-berri in https://github.com/BerriAI/litellm/pull/25924

## New Contributors
* @shreyescodes made their first contribution in https://github.com/BerriAI/litellm/pull/25559
* @Lucas-Song-Dev made their first contribution in https://github.com/BerriAI/litellm/pull/25324

**Full Changelog**: https://github.com/BerriAI/litellm/compare/v1.83.8-nightly...v1.83.9-nightly