What's Changed
- fix bedrock cost calculation for cached tokens by @jdietzsch91 in #12488
- Fix tool call handling in Anthropic pass-through adapter by @iwinux in #12473
- Guardrails AI - pre-call + logging only guardrail (pii detection/competitor names) support by @krrishdholakia in #12506
- Litellm mcp access group on UI by @jugaldb in #12470
- [Enterprise] Support tag based mode for guardrails by @krrishdholakia in #12508
- Litellm mcp access group by @jugaldb in #12514
- Add
Build and push litellm-non_root
todocker-hub-deploy
CICD workflow by @andresC98 in #12413 - Validation to mcp server name by @jugaldb in #12515
- [Feat] - New guardrail - OpenAI Moderations API by @ishaan-jaff in #12519
- [MCP Gateway] QA - MCP Tool Testing Playground by @ishaan-jaff in #12520
- [Security Fix] - Dont show pure JWT in "Logs" page on UI by @ishaan-jaff in #12524
- [Bug Fix] - QA for MCP Gateway - show the cost config on the root of MCP Settings by @ishaan-jaff in #12526
- [MCP Gateway] access group UI object permission fix by @jugaldb in #12523
- [MCP Gateway] UI Quality check fixes by @jugaldb in #12521
- [MCP Gateway] Allow using stdio MCPs with LiteLLM by @ishaan-jaff in #12530
- docs: Update github.md by @EmaSuriano in #12509
- 🐛 Remove deprecated pydantic class Config by @strawgate in #12528
- Team Members - reset budget, if duration set + Prometheus - support tag based metrics by @krrishdholakia in #12534
- Consistent layout for Create and Back buttons on all the pages by @NANDINI-star in #12542
- Fix e2e test by @NANDINI-star in #12544
- Align Show Password with Checkbox by @NANDINI-star in #12538
- chore: Update Vertex AI Model Garden LiteLLM integration tutorial by @lizzij in #12428
- [Bug Fix] xai/ translation fix - ensure finish_reason includes tool calls when xai responses with tool calls by @ishaan-jaff in #12545
- Prevent writing default user setting updates to yaml (error in non-root env) + Use central team member budget when max_budget_in_team set on UI by @krrishdholakia in #12533
- [MCP Gateway] Allow mcp access groups on test key and tool calls by @jugaldb in #12529
- [MCP Gateway] UI headers groups example on connect tab by @jugaldb in #12550
- Fix e2e test by @NANDINI-star in #12549
- Integration: Bytez as a model provider by @inf3rnus in #12121
New Contributors
- @jdietzsch91 made their first contribution in #12488
- @iwinux made their first contribution in #12473
- @andresC98 made their first contribution in #12413
- @EmaSuriano made their first contribution in #12509
- @strawgate made their first contribution in #12528
- @inf3rnus made their first contribution in #12121
Full Changelog: v1.74.2-nightly...v1.74.3-nightly
Docker Run LiteLLM Proxy
docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
ghcr.io/berriai/litellm:main-v1.74.3-nightly
Don't want to maintain your internal proxy? get in touch 🎉
Hosted Proxy Alpha: https://calendly.com/d/4mp-gd3-k5k/litellm-1-1-onboarding-chat
Load Test LiteLLM Proxy Results
Name | Status | Median Response Time (ms) | Average Response Time (ms) | Requests/s | Failures/s | Request Count | Failure Count | Min Response Time (ms) | Max Response Time (ms) |
---|---|---|---|---|---|---|---|---|---|
/chat/completions | Failed ❌ | 210.0 | 397.6239629079091 | 5.915434271857456 | 0.0 | 1770 | 0 | 185.11648200001218 | 16174.709219000022 |
Aggregated | Failed ❌ | 210.0 | 397.6239629079091 | 5.915434271857456 | 0.0 | 1770 | 0 | 185.11648200001218 | 16174.709219000022 |