v1.107.0-dev.2Pre-release
Added 6
- Added installationPreference emission in marketplace.json entries for Claude Code
- Added integration test coverage for fine_tuning endpoint contracts
- Auto mode safeguard_results are generated with a classifier model on the /v1/messages chat bridge
- Separate confidential client for ID-JAG leg 2 in MCP
- Added Laya and Bespoke providers to the Add Model form
- Added Wan 3.0, Wan 2.7 and HappyHorse video generation via /v1/videos for DashScope
Changed 4
- Decision models are shown as decisions and the supported ones are listed in the AI Hub
- Images are sent to Cloudflare Clef through its images field
- Replaced kerrylu-berri with kerry-berri in CODEOWNERS
- Configured logging callbacks are built at boot so a broken callback fails startup instead of the first request
Fixed 10
- Scoped /global/activity/cache_hits analytics to the caller's own and permitted-team logs
- Cached tool result screenshots on the /v1/messages to Responses bridge
- Databricks now sends a single plain text block as a string for non-Claude models
- POST and PATCH are only replayed when the request never reached the upstream
- Widened the LiteAdmin home model selector
- Cache hits kept free of unset fields and CircleCI suites updated to follow intentional behavior changes
- Vertex forced-tool support is checked on the real model, non-string user ids are skipped, and speed is dropped where unsupported
- Registry audit removed azure_ai/claude-fable-5-5 from the cost map
- service_tier is forwarded on chat completions to Anthropic and Azure
- Decisions playground issues are labeled by question key and the Response header row is fixed
Security 1
- Empty identities are rejected before session creation in SSO
From LiteLLM
Verify Docker Image Signature
All LiteLLM Docker images are signed with cosign. Every release is signed with the same key introduced in commit 0112e53.
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:
cosign verify \
--key https://raw.githubusercontent.com/BerriAI/litellm/0112e53046018d726492c814b3644b7d376029d0/cosign.pub \
ghcr.io/berriai/litellm:v1.107.0-dev.2
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:
cosign verify \
--key https://raw.githubusercontent.com/BerriAI/litellm/v1.107.0-dev.2/cosign.pub \
ghcr.io/berriai/litellm:v1.107.0-dev.2
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
- feat(claude-code): emit installationPreference in marketplace.json entries by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/41430
- fix(proxy): scope /global/activity/cache_hits analytics to the caller's own and permitted-team logs by @ojensen-berri in https://github.com/BerriAI/litellm/pull/46210
- fix(anthropic): cache tool result screenshots on the /v1/messages to Responses bridge by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46324
- test(integration): cover fine_tuning endpoint contracts by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/44697
- fix(databricks): send a single plain text block as a string for non-Claude models by @mrinal-berri in https://github.com/BerriAI/litellm/pull/45686
- feat(ui): show decision models as decisions and list the supported ones in the AI Hub by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45965
- feat(decisions)!: send images to Cloudflare Clef through its images field by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45734
- feat(anthropic): generate auto mode safeguard_results with a classifier model on the /v1/messages chat bridge by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45600
- feat(mcp): separate confidential client for ID-JAG leg 2 by @joshua-berri in https://github.com/BerriAI/litellm/pull/46211
- fix(http_handler): only replay POST and PATCH when the request never reached the upstream by @mateo-berri in https://github.com/BerriAI/litellm/pull/46336
- feat(decisions): add Laya and Bespoke providers to the Add Model form by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45260
- fix(ui): widen the LiteAdmin home model selector by @moe-berri in https://github.com/BerriAI/litellm/pull/46446
- fix(cache): keep cache hits free of unset fields, and follow intentional behavior changes in CircleCI suites by @mateo-berri in https://github.com/BerriAI/litellm/pull/46359
- fix(anthropic): check Vertex forced-tool support on the real model, skip non-string user ids, drop speed where unsupported by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46337
- fix(sso)!: reject empty identities before session creation by @yucheng-berri in https://github.com/BerriAI/litellm/pull/44098
- chore(codeowners): replace kerrylu-berri with kerry-berri by @kerry-berri in https://github.com/BerriAI/litellm/pull/46452
- fix(cost-map): registry audit 2026-10-14, remove azure_ai/claude-fable-5-5 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46428
- fix(proxy): build configured logging callbacks at boot so a broken callback fails startup instead of the first request by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46378
- feat(dashscope): add Wan 3.0, Wan 2.7 and HappyHorse video generation via /v1/videos by @Bytechoreographer in https://github.com/BerriAI/litellm/pull/43706
- fix(anthropic): forward service_tier on chat completions to Anthropic and Azure by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46461
- fix(ui): label Decisions playground issues by question key and fix the Response header row by @mateo-berri in https://github.com/BerriAI/litellm/pull/46145
- fix(http): restore the 15s aiohttp keep-alive default in the Python and Rust clients by @mateo-berri in https://github.com/BerriAI/litellm/pull/46355
- fix(vertex_ai): pass the per-request timeout to Gemini streaming calls by @kerry-berri in https://github.com/BerriAI/litellm/pull/46349
- feat(proxy): add deployment warning endpoints with instance-wide ignore and snooze by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46347
- fix(proxy): check previous_response_id ownership on /v1/responses/compact by @mateo-berri in https://github.com/BerriAI/litellm/pull/46408
- refactor(router): name the pre-content stream retry helpers for both stream paths and let the retry loop take its attempt function by @kerry-berri in https://github.com/BerriAI/litellm/pull/46474
- fix(anthropic): stream json_schema answers as content on Bedrock Mantle and Claude Platform, and join split answers by @mateo-berri in https://github.com/BerriAI/litellm/pull/46331
- fix(caching): collect the Redis circuit-breaker state gauge as one series per state across worker processes by @kerry-berri in https://github.com/BerriAI/litellm/pull/46357
- fix(gemini): add the changelog shutdown date to gemini-2.0-flash-exp-image-generation by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/46489
- test(integration): remove six setup causes of CircleCI integration reds by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46490
- fix(s3_v2): bound each upload attempt, add jitter, retry 429 and transport errors in the call by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46340
- fix(streaming): add CustomStreamWrapper.close() and release the connection when a sync stream ends by @kerry-berri in https://github.com/BerriAI/litellm/pull/46354
- feat(proxy): let proxy admins open soft_budget to team admins by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46485
- fix(proxy): accept every known logging callback on /health/services by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46297
- test(proxy): cover the OTel V2 otel callback on /health/services by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46394
- fix(vllm): write spend log rows for /vllm passthrough calls to router models by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45733
- fix(config): treat blank env values as unset so the proxy neither crash-loops nor silently drops a guardrail by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46341
- fix(proxy): evict the cached user on every proxy for every /user/update and /user/bulk_update by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46500
- feat(proxy): let proxy admins open team member defaults to team admins by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46353
- fix(anthropic): forward service_tier on /v1/messages for anthropic, azure_ai and bedrock_mantle by @kerry-berri in https://github.com/BerriAI/litellm/pull/46370
- fix(router): read the first stream chunk before returning so a failure before any token runs through the retry and fallback budget by @kerry-berri in https://github.com/BerriAI/litellm/pull/46493
- fix(caching): report Redis circuit-breaker metrics through the Prometheus services logger by @kerry-berri in https://github.com/BerriAI/litellm/pull/46484
- fix(anthropic): keep refusal stop_reason and stop_details on the logged response by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46104
- fix(router): add fallback_on_unsupported_params so param rejections can raise instead of falling back by @kerry-berri in https://github.com/BerriAI/litellm/pull/46350
- feat(access-groups): per-server MCP tool permissions on access groups by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46300
- feat(helm): read database and Redis host or URL from a Kubernetes Secret by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46410
- fix(mcp): leave out tool texts too long to embed instead of failing tool search by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/43554
- perf(scim): batch group member identity lookups into one query per push by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46412
- feat(helm): support a separate hostname per component in the componentized deploy by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46360
- fix(redis): default every Redis client to a 5s socket timeout when none is set by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46501
- fix(router): report fallbacks skipped for access or budget instead of calling them failed by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46291
- fix(proxy): finish prisma engine requests when their caller is cancelled by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46503
- fix(proxy-extras): run prisma migrate deploy under the pinned coordinator lock instead of Prisma's session advisory lock by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46368
- test(integration): fix three more CircleCI integration reds from test setup by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46510
- fix(caching): make Redis Cluster commands wait for a pool slot instead of tripping the circuit breaker by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46488
- test(rust): assert native serving via the x-litellm-rust response marker by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46531
- fix(responses): propagate cache_control through responses->chat completion bridge by @TobiMayr in https://github.com/BerriAI/litellm/pull/32511
- feat(guardrails): add Airia guardrail integration by @airia-open-source in https://github.com/BerriAI/litellm/pull/40784
- feat(terraform): add team_member_model_max_budget to litellm_team by @matthowardcohere in https://github.com/BerriAI/litellm/pull/46274
- fix(responses): keep the provider code and message on top-level stream error events by @mateo-berri in https://github.com/BerriAI/litellm/pull/46438
- feat(access-groups): grant Claude Code skills through access groups by @yassin-berriai in https://github.com/BerriAI/litellm/pull/46294
- test(integration): delete flaky team model TPM concurrency test by @kerry-berri in https://github.com/BerriAI/litellm/pull/46541
- fix(vertex-ai): add flex prices to gemini-nano-banana-2.1 by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/46512
- feat(proxy): let proxy admins open team metadata to team admins, keeping proxy settings with proxy admins by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46506
- fix(responses): return the first message item as content for structured output on the chat bridge by @kerry-berri in https://github.com/BerriAI/litellm/pull/46508
- fix(cost): split xai spend rows across input and output instead of booking the total as output by @kerry-berri in https://github.com/BerriAI/litellm/pull/46320
- test(rust-bridge): pin Messages and Responses callback contracts on the native path by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46443
- test(host-python): move driver tests out of the inline module by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46325
- refactor(host-python): collect host failures in a thiserror Error enum by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46361
- refactor(host-python): run the Python driver as one async body by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46235
- refactor(rust-bridge): generator-style execution protocol with stream step names by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46282
- test(translation): share rich request and response templates by @kerry-berri in https://github.com/BerriAI/litellm/pull/46546
- refactor(rust-bridge): return an unstarted Execution from native entrypoints and drive it in Python by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/46299
- test(integration): add claude-haiku-5-5 basic translation cases by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45176
- feat(ui): add the Airia guardrail to the Guardrail Garden with its logo by @airia-open-source in https://github.com/BerriAI/litellm/pull/46562
- feat(proxy): return the W3C traceresponse header on traced responses by @adityajoy-1902 in https://github.com/BerriAI/litellm/pull/42079
- feat(proxy): general_settings.disable_requester_ip_address_logging keeps caller IP out of spend logs by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45804
- test(integration): update stale CircleCI integration expectations by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46570
- fix(cli): run group subcommands in the lite shell and default to the logged-in gateway by @mateo-berri in https://github.com/BerriAI/litellm/pull/46511
- fix(cost): price AgentCore web search at $7 per 1,000 queries by @litellm-stampbot in https://github.com/BerriAI/litellm/pull/46571
- feat(ui): replace deployment warning banners with a health icon, alert banner and panel by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46498
- fix(caching): report zero response_cost on a response-cache hit by @Apeksha9389 in https://github.com/BerriAI/litellm/pull/46471
- fix(responses bridge): keep defer_loading on function tools sent to the Responses API by @mateo-berri in https://github.com/BerriAI/litellm/pull/46439
- fix(router): read the first stream chunk through aiter so MCP auto-run streaming works again by @kerry-berri in https://github.com/BerriAI/litellm/pull/46578
- fix(anthropic): keep tool_search_tool_result so the search call replays with its result by @mateo-berri in https://github.com/BerriAI/litellm/pull/46440
- test(translation): add gpt-6-astra translation cases by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46580
- test(translation): recapture bedrock invoke claude-sonnet-4-6 bases after usage shape change by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46579
- test(translation): add claude-fable-5-1 translation cases by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46581
- fix(bedrock)!: replay redacted_thinking on Converse and keep its data as Claude's own string by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46532
- test(translation): share rich fixtures across Claude, GPT and Gemini providers by @kerry-berri in https://github.com/BerriAI/litellm/pull/46585
- fix(bedrock): use base_model for Claude capability checks on application inference profile ARNs by @ryan-crabbe-berri in https://github.com/BerriAI/litellm/pull/46569
- feat(cost-map): add scx-ai DeepSeek-V4.1-flash, GLM-5.3 and Kimi-K3 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46496
- fix(cost): price typesafe/jev-1.14.0 by @hyamie in https://github.com/BerriAI/litellm/pull/46594
- fix(responses): translate OpenAI tool_search to Anthropic tool search for Claude models by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45882
- test(integration): follow #46101 in the breakpoint cache mark census expectation by @mateo-berri in https://github.com/BerriAI/litellm/pull/46509
- fix(router): keep the prompt caching affinity pin alive as long as the prefix's cache_control ttl by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46084
- fix(anthropic): emit one tool_use block per call when a /v1/messages stream chunk carries several by @narmstrong-berri in https://github.com/BerriAI/litellm/pull/46492
- refactor: clean up fresh tech debt in per-user Copilot auth, Ternary FOCUS export and telemetry by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45769
- fix(responses bridge): round-trip tool_search_call and tool_search_output items between turns by @mateo-berri in https://github.com/BerriAI/litellm/pull/46444
- fix(anthropic)!: translate the OpenAI tool_search tool on /v1/chat/completions for Claude models by @mateo-berri in https://github.com/BerriAI/litellm/pull/46389
- fix(xai): name a completed xAI batch's results file and stop looking it up on xAI's files API by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45559
- chore(prices): add xAI grok-voice-3.0, grok-voice-2.0 and grok-voice-latest realtime rows by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/46602
- fix(responses): map Anthropic tool search to OpenAI server tool search by @mateo-berri in https://github.com/BerriAI/litellm/pull/46441
- fix(proxy): store list-valued audit log changes as JSON text by @yuneng-berri in https://github.com/BerriAI/litellm/pull/46555
New Contributors
- @airia-open-source made their first contribution in https://github.com/BerriAI/litellm/pull/40784
- @matthowardcohere made their first contribution in https://github.com/BerriAI/litellm/pull/46274
- @adityajoy-1902 made their first contribution in https://github.com/BerriAI/litellm/pull/42079
- @Apeksha9389 made their first contribution in https://github.com/BerriAI/litellm/pull/46471
- @hyamie made their first contribution in https://github.com/BerriAI/litellm/pull/46594
- @narmstrong-berri made their first contribution in https://github.com/BerriAI/litellm/pull/46492
Full Changelog: https://github.com/BerriAI/litellm/compare/v1.107.0-dev.1...v1.107.0-dev.2