BerriAI/litellm v1.106.0 on GitHub

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:v1.106.0

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/v1.106.0/cosign.pub \
  ghcr.io/berriai/litellm:v1.106.0

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

test(mcp): fire MCP client test deadlines on conditions instead of wall-clock time by @devin-ai-integration[bot] in #44166
fix(cost-map): add OpenAI TTS and GPT-5.x deprecation dates by @devin-ai-integration[bot] in #44175
fix(tracing): unify ClickHouse storage configuration by @yujonglee-berri in #43941
feat(docker): one-command quickstart that starts the gateway, Postgres, and the admin UI by @misbahsy in #43673
fix(lens): simplify the example investigation preview by @moe-berri in #44123
chore(helm): drop migrationJob values the chart never reads by @Bernedotcom2312 in #42141
fix(ui): register tencent in the Add Model provider dropdown by @jimaldon in #40924
test: remove 130 legacy tests owned by stronger unit proofs by @devin-ai-integration[bot] in #44157
feat(mcp)!: disable stdio MCP servers by default by @yuneng-berri in #44066
fix(proxy): exit when DATABASE_URL is set but the Prisma toolchain is missing by @devin-ai-integration[bot] in #44207
fix(proxy): preserve lifespan state across all entrypoints by @tin-berri in #44214
fix(oci): resolve the GenAI endpoint realm from the compartment OCID instead of hardcoding oraclecloud.com by @fede-kamel in #43180
fix(chatgpt): preserve requested service tier in Responses calls by @SteffanWolter in #40108
fix(ui): leave unset callback select params out of the save payload by @devin-ai-integration[bot] in #44213
fix(proxy-extras): hand libpq a root cert, not Prisma's sslcert, when the migration job builds indexes by @devin-ai-integration[bot] in #44203
feat: add Laya gateway and OSS classifier providers by @tin-berri in #43626
fix(auto-router): count usage savings by selected UTC request day by @tin-berri in #44115
fix(lens): preserve framework agent names and GenAI message content by @moe-berri in #44218
feat(jwt): auto_register_map_existing_key maps JWT to the user's existing virtual key by @devin-ai-integration[bot] in #42375
feat(ui): select Laya for OSS classification by @tin-berri in #43768
fix(otel): tolerate non-dict callback_settings.otel and ignore bare EXCLUDED_SERVICES env by @devin-ai-integration[bot] in #44086
fix(proxy): carry key, team and project tags into pass-through spend logs by @devin-ai-integration[bot] in #42662
feat(ui): add System One (Jev) tab to the playground by @devin-ai-integration[bot] in #44043
chore: bump litellm-proxy-extras 0.4.104 -> 0.4.105 by @devin-ai-integration[bot] in #44234
chore(prices): add xAI grok-voice-transcribe-1.0 deprecation date by @berriai-litellm-provider-info-sync[bot] in #44237
fix(exa): fall back to highlights/summary when text is missing by @wakqasahmed in #42213
fix(scim): apply path-less group PATCH ops instead of storing them under an empty metadata key by @devin-ai-integration[bot] in #43978
fix(lens): run investigations with configured wildcard models by @moe-berri in #44233
fix(ui): shrink the sidebar logo so it stops outweighing page titles by @yuneng-berri in #44247
feat(tracing): support claude agent sdk traces with agent name, logo and chat content by @ishaan-berri in #44248
fix(ui): make model leaderboard chart bars wide and readable by @devin-ai-integration[bot] in #44249
test: repair stale and polluting tests red on scheduled main CI by @yuneng-berri in #44229
feat(traces): type queries and align read access with log visibility by @yujonglee-berri in #44228
chore(ui): untrack tsconfig.tsbuildinfo and gitignore *.tsbuildinfo by @devin-ai-integration[bot] in #44261
refactor(mcp): extract upstream preparation and support modern clients by @joshua-berri in #44232
fix(lens): default to traces and reopen span details by @moe-berri in #44260
fix(bedrock): surface unrecognized converse-stream event frames instead of an empty turn by @devin-ai-integration[bot] in #44112
perf(proxy): batch daily model usage writes instead of upserting per request by @ishaan-berri in #44243
test(integration): opt the config pass-through spend-log case into auth by @yuneng-berri in #44265
test: fix three order-dependent and timing-flaky tests by @yuneng-berri in #44271
fix(ui): make the Lens traces refresh button always clickable by @devin-ai-integration[bot] in #44252
feat(lens): add sample previews and improve setup and worker feedback by @moe-berri in #44268
chore(openrouter): sync prices, limits and deprecation dates from the models API by @berriai-litellm-provider-info-sync[bot] in #44287
refactor(traces): type the ClickHouse query help response by @devin-ai-integration[bot] in #44285
fix(proxy): stop queued registry read-throughs spending the resync budget by @yuneng-berri in #44277
feat(lens): show investigation progress as one staged bar with time left by @ishaan-berri in #44301
refactor(ui): compose dashboard pages with shared layouts by @devin-ai-integration[bot] in #44306
fix(proxy): share ownership permissions for spend logs and traces by @yujonglee-berri in #44239
fix(proxy-extras): retry P3009 when a peer already recovered the named migration row by @devin-ai-integration[bot] in #44283
fix(bedrock): price amazon nova 2 pro preview at the standard tier by @berriai-litellm-provider-info-sync[bot] in #44302
feat: add Bespoke Nimble gateway and OSS classifier support by @tin-berri in #44246
fix(anthropic): stop repeating streamed thinking text in the signature chunk by @devin-ai-integration[bot] in #44127
fix(responses): drop bridge-minted reasoning items from OpenAI replays by @devin-ai-integration[bot] in #44132
feat(bedrock): drop lookaround regex patterns from tool schemas for Converse models that reject them by @devin-ai-integration[bot] in #44138
feat(bedrock): serve gpt-5.6+ chat completions natively by default, with chat_completions/ opt-in for gpt-oss and grok by @devin-ai-integration[bot] in #44307
feat(lens): issue briefs with problem, user goal, outcome and test cases by @ishaan-berri in #44311
fix(responses): keep gpt-5.4/5.5 tool calls on chat and merge bridged tool calls into one choice by @devin-ai-integration[bot] in #44295
fix(cost): price rule-only model names at the deployment's rate by @devin-ai-integration[bot] in #44144
fix(auto-router): attribute router-day savings without spend logs or a session by @tin-berri in #44226
test(integration): let run.py select cells by pytest node id by @devin-ai-integration[bot] in #44319
fix(bedrock): honor unsupported reasoning effort levels for OpenAI GPT models by @Aasif-Multani in #44183
feat(lens): add preset watch-for checks to investigation setup by @ishaan-berri in #44313
fix(proxy): use OpenAI workload identity federation tokens on /openai_passthrough by @6matt in #44140
feat(interactions): durable cross-pod settlement for background interaction billing by @devin-ai-integration[bot] in #41955
fix(streaming): keep the finish_reason of the provider's last content chunk in the logged response by @joeym82956 in #44318
fix(bedrock): send the anthropic-workspace-id header to Bedrock Mantle by @kohjunhao in #44173
fix(guardrails): run unified guardrails on /v1/images/edits by @michelligabriele in #44195
feat(scaleway): add rerank support by @ankit373 in #44160
fix(vector_stores): return managed file ids from vector store file list by @devin-ai-integration[bot] in #43800
fix(proxy): return 4xx instead of 500 for missing required params, invalid pagination and unknown ids by @devin-ai-integration[bot] in #43787
fix(guardrails): encrypt guardrail litellm_params secrets at rest by @yucheng-berri in #43627
fix(search_tools): encrypt search tool litellm_params at rest by @yucheng-berri in #43631
feat(roi): add GitLab sources and branch cost attribution by @moe-berri in #44324
fix(model-prices): mark azure us/eu responses-only models as mode responses by @devin-ai-integration[bot] in #44323
revert(cost): revert "fix(cost): price rule-only model names at the deployment's rate" (#44144) by @devin-ai-integration[bot] in #44335
fix(proxy): stamp the client alias on a copy of each streamed chunk so pricing sees the deployment model by @devin-ai-integration[bot] in #44341
revert(responses): revert "fix(responses): keep gpt-5.4/5.5 tool calls on chat and merge bridged tool calls into one choice" (#44295) by @devin-ai-integration[bot] in #44344
refactor(dashboard): migrate to zod 4 and openai 6 by @devin-ai-integration[bot] in #44345
feat(ui): move LiteAdmin into the header with a docked side panel by @tin-berri in #44293
fix(responses): merge bridged tool calls into the same choice as the text by @devin-ai-integration[bot] in #44346
feat(ui): improve trace inspection and ROI estimation by @moe-berri in #44351
refactor(ui): reorganize Lens dashboard components by @devin-ai-integration[bot] in #44354
refactor(ui): inject Lens backends as services instead of a fake HTTP client by @devin-ai-integration[bot] in #44369
refactor: clean up fresh tech debt from 2026-10-02 by @devin-ai-integration[bot] in #44362
feat(traces): tracing development seed by @devin-ai-integration[bot] in #44363
refactor(types): replace Any with proven types in 4 files by @devin-ai-integration[bot] in #44370
build(deps): suppress unfixed braces GHSA-vfj7-8cjw-p6xm to clear osv-scan by @devin-ai-integration[bot] in #44347
fix(azure): add MAI-Image max_input_tokens from models sold directly page by @berriai-litellm-provider-info-sync[bot] in #44375
fix(proxy): emit postgres service spans only on real DB reads in auth cache helpers by @devin-ai-integration[bot] in #44148
fix(otel): nest cache spans under their operation and name service spans by purpose by @devin-ai-integration[bot] in #44150
fix(otel): name postgres service spans by operation and table by @devin-ai-integration[bot] in #44240
perf(proxy): stop prompt-cache eligibility from tokenizing the whole conversation by @devin-ai-integration[bot] in #44221
fix(proxy): keep the database error when the log_db_metrics failure hook raises by @devin-ai-integration[bot] in #44383
feat(azure): add azure_ai/kimi-k2-thinking from Azure Kimi pricing page by @berriai-litellm-provider-info-sync[bot] in #44382
feat(harness): add Harness.TOOL_LOOP, a minimal in-process tool-calling loop by @devin-ai-integration[bot] in #44391
feat(decisions): add unified /v1/decisions endpoint for Jev-compatible providers by @devin-ai-integration[bot] in #44236
fix(lens): paginate trace reads within ClickHouse limits by @moe-berri in #44384
feat(lens): live dot field timeline and full-screen traces view by @ishaan-berri in #44390
feat(ui): make LiteAdmin enterprise-only by @tin-berri in #44399
fix(bedrock): stop emitting Converse cachePoint blocks for Kimi K3 by @devin-ai-integration[bot] in #44292
test(cost): pin OpenAI reported web search count with mixed actions by @devin-ai-integration[bot] in #44414
fix(ci): give the pass-through auth regression test a real FastAPI app by @devin-ai-integration[bot] in #44403
fix(azure): add azure_ai/flux.2-pro input token limit from the models sold directly page by @berriai-litellm-provider-info-sync[bot] in #44405
fix(lens): batch run reads and reset trace pagination by @moe-berri in #44398
fix(lens): preserve full trace access and expose investigation failures by @moe-berri in #44406
feat: add make lens-dev for one-command lens local dev by @ishaan-berri in #44413
fix(bedrock): honor the per-request timeout on Converse and Invoke streaming (internal copy of #38210) by @devin-ai-integration[bot] in #44134
feat(traces): render curated SQL examples from shared files by @devin-ai-integration[bot] in #44388
fix(model_prices): registry audit 2026-10-03, add cohere embed v5, grok-imagine-video-1.5-lite and openrouter gpt-image rows by @devin-ai-integration[bot] in #44376
refactor(traces): extract snapshot cache by @devin-ai-integration[bot] in #44424
feat(lens): always-on investigations with findings and investigations tables by @ishaan-berri in #44418
feat(docker): add a Windows quickstart in PowerShell, and a styled terminal for both quickstarts by @misbahsy in #44310
feat(proxy): serve a Codex-native model catalog with per-model service_tiers from /v1/models by @devin-ai-integration[bot] in #44136
refactor(types): replace Any with proven types in 9 files by @devin-ai-integration[bot] in #44389
fix(proxy): treat Postgres connection exhaustion as DB unavailable, not poison spend-log rows by @devin-ai-integration[bot] in #44270
fix(proxy): treat Postgres connection exhaustion as backpressure, not poison rows by @devin-ai-integration[bot] in #44266
fix(lens): release budget reservations when the analysis model call fails by @devin-ai-integration[bot] in #44431
fix(mcp): preserve upstream tool schemas and parameter headers by @joshua-berri in #44425
fix(bedrock_mantle): route Claude chat completions to the native Messages endpoint by @devin-ai-integration[bot] in #43646
fix(tracing): preserve spend identity and gateway correlation by @yujonglee-berri in #44421
feat(roi): measure shipping velocity, quality, and recorded spend by @moe-berri in #44426
fix(azure): set text-embedding max input to 8192 from the models sold directly page by @berriai-litellm-provider-info-sync[bot] in #44449
feat(ui): show invitation and reset password links in a copyable field by @yuneng-berri in #44454
feat(lens): coordinate worker releases and bundled installs by @moe-berri in #44428
fix(auth): resolve hidden model_group_alias entries in the zero-cost budget check by @devin-ai-integration[bot] in #43741
fix(vertex_ai): forward system and tools to partner model count_tokens by @devin-ai-integration[bot] in #43900
fix(traces): reject conflicting spend aliases and unrelated HTTP siblings by @devin-ai-integration[bot] in #44456
chore(cost-map): add azure_ai/kimi-k2-thinking retirement date from the Azure retired models page by @berriai-litellm-provider-info-sync[bot] in #44455
fix(health): probe Bedrock Mantle Claude deployments over the Anthropic Messages API by @devin-ai-integration[bot] in #44419
fix(tests): match the OS bind error in the owned-proxy port-race retry by @devin-ai-integration[bot] in #44462
feat(anthropic): workload identity federation and pluggable identity sources by @devin-ai-integration[bot] in #44448
test(integration): exact four-part translation cases on a shared fake provider and shared YAML deployment by @devin-ai-integration[bot] in #44451
feat(roi): default people and branch lists to matched accounts by @moe-berri in #44465
feat(enterprise): bundle LiteAdmin Slack with native gateway login by @tin-berri in #44444
fix(caching): count tool_call cache_control marks in the injection census by @devin-ai-integration[bot] in #43556
chore(cost-map): sync openrouter prices from the models API by @berriai-litellm-provider-info-sync[bot] in #44466
fix(lens): preserve approved worker digests and harden its image by @moe-berri in #44467
feat: improve Lens dev seeding and live UI by @devin-ai-integration[bot] in #44468
test(ui): wait for step search value to settle in TraceDrawer test by @devin-ai-integration[bot] in #44470
feat(ui): share trace drawer as a closable SidePanel and polish Lens by @devin-ai-integration[bot] in #44473
fix(lens): align source setup with available worker images by @moe-berri in #44476
feat(sdk): add run_tool_loop and arun_tool_loop helpers by @devin-ai-integration[bot] in #44381
feat(lens): guide setup through the first investigation by @moe-berri in #44475
test(ci): fix six CircleCI test regressions on main by @devin-ai-integration[bot] in #44429
fix(proxy-extras): log v1 migration failures at ERROR so LITELLM_LOG=ERROR shows them by @devin-ai-integration[bot] in #44202
feat(ui): inline Lens settings tab and investigation editor by @devin-ai-integration[bot] in #44479
feat(mcp): hand listed-tool description and input schema to pre-call hooks per caller by @devin-ai-integration[bot] in #41162
refactor: clean up fresh tech debt from 2026-10-03 by @devin-ai-integration[bot] in #44484
fix(otel): read the registered v2 logger without importing the proxy by @devin-ai-integration[bot] in #44485
test(mcp): keep the SSO assertion round trip from matching its refresh token inside random ciphertext by @devin-ai-integration[bot] in #44494
test(ci): pin the ROI estimator flag and the prompt-cache counter in two drifted tests by @devin-ai-integration[bot] in #44499
refactor(ui): move TraceView into components/lens/traces by @devin-ai-integration[bot] in #44501
refactor(ui): rename view_logs to logs and split request, audit and detail by @devin-ai-integration[bot] in #44505
refactor(ui): share CopyButton between Lens traces and logs by @devin-ai-integration[bot] in #44513
refactor(lens): storage-independent trace reads, shared keyset pager, typed read failures by @devin-ai-integration[bot] in #44422
fix(ci): send the Lens preview body as selection in the tracing endpoint tests by @devin-ai-integration[bot] in #44525
fix(azure): set gpt-realtime-2.1 retirement dates from the retirement schedule by @berriai-litellm-provider-info-sync[bot] in #44526
fix(ci): repair the security sweep and Lens billing integration tests for ROI, JWKS, and release identity changes by @devin-ai-integration[bot] in #44529
fix(dashboard): route dev API calls on Accept and fail fast on non-JSON 2xx by @devin-ai-integration[bot] in #44528
chore(cost-map): sync openrouter prices from the models API by @berriai-litellm-provider-info-sync[bot] in #44533
refactor(ui): route dashboard URL state through nuqs parsers by @devin-ai-integration[bot] in #44537
fix(logging): deduplicate streaming failure callbacks by @devin-ai-integration[bot] in #44442
feat(proxy): add LITELLM_FIPS_MODE startup gate with provider assertion and loud password migration failure by @devin-ai-integration[bot] in #42700
fix(guardrails): run the end-of-stream post_call scan when the client disconnects mid-stream by @devin-ai-integration[bot] in #43839
refactor(mcp): fix Sequence/list return mismatch and collapse record_listed_tools wrapper by @devin-ai-integration[bot] in #44556
chore(cost-map): add azure retirement dates for gpt-4o-realtime-preview-2024-10-01 and jamba-instruct by @berriai-litellm-provider-info-sync[bot] in #44567
feat(ui): add persistent columns and loading skeletons to Lens runs by @devin-ai-integration[bot] in #44579
fix(ui): remove Top models by task card from Model Leaderboard by @devin-ai-integration[bot] in #44502
refactor(rust): track ClickHouse migrations in a checksummed ledger by @devin-ai-integration[bot] in #44580
refactor(ui): extract shared timeline and time-range controls by @devin-ai-integration[bot] in #44584
fix(azure): set gpt-4o-transcribe retirement date from the retirement schedule by @berriai-litellm-provider-info-sync[bot] in #44585
ci: move Postgres, MCP and Redis suites to CircleCI integration by @devin-ai-integration[bot] in #44453
chore!: retire the integrated ROI calculator by @moe-berri in #44477
fix(caching): key response cache by router model group in litellm_metadata by @devin-ai-integration[bot] in #44542
fix(auto-router): align preview and serving default resolution by @moyai-berriai[bot] in #44436
fix(mcp): bind OAuth clients to their upstream issuer by @devin-ai-integration[bot] in #37777
feat(guardrails): add llm shield pii redaction and rehydration guardrail by @ninadphalak in #42645
refactor(tracing): generate existing HTTP request models from Rust schemas by @devin-ai-integration[bot] in #44591
ci: run unit selections from GHA test-path and drop the CircleCI unit jobs by @devin-ai-integration[bot] in #44461
fix(ui): make mobile sidebar and top bar responsive by @moe-berri in #44603
refactor(types): replace Any with proven types in 137 files by @mateo-berri in #44478
fix(ui): restore inline Lens onboarding and responsive layout by @moe-berri in #44604
test(integration): anthropic-route basic translation cases for six Claude models on messages, chat completions and responses by @devin-ai-integration[bot] in #44607
fix(responses): report truncated bridged output as incomplete and echo request params by @devin-ai-integration[bot] in #44460
fix(proxy): enforce internal-user model creation prohibition by @moyai-berriai[bot] in #44438
chore(harness): drop the unused Any import left in harness options by @devin-ai-integration[bot] in #44625
fix(harness): drop the unused Any import that fails ruff on main by @devin-ai-integration[bot] in #44627
refactor(proxy): rename management/teams/access.py to authz.py by @ryan-crabbe-berri in #44624
feat(tracing)!: return only data from SQL queries by @devin-ai-integration[bot] in #44609
fix(openai/realtime): drop model from upstream URL for intent=transcription by @michelligabriele in #43854
fix(gemini): add priority tier audio input price to three Gemini rows by @berriai-litellm-provider-info-sync[bot] in #44632
fix(cost): bill per-second transcription models outside chat modes by @devin-ai-integration[bot] in #44458
fix(ui): explain why team member reset spend is unavailable instead of hiding it by @devin-ai-integration[bot] in #44629
test(integration): bedrock_converse-route basic translation cases for six Claude models on messages, chat completions and responses by @devin-ai-integration[bot] in #44646
test(integration): run the anthropic /v1/responses basic translation cases now that the bridge echoes request params by @devin-ai-integration[bot] in #44615
feat(lens): restore compact navigation with live investigation review by @ishaan-berri in #44472
feat(proxy): limit which models an end user can call by @devin-ai-integration[bot] in #43904
fix(sso): let CLI and Claude Code gateway sign-in through on DISABLE_ADMIN_UI nodes by @devin-ai-integration[bot] in #44620
feat(ui): declare shared search operators and flush queries on blur by @devin-ai-integration[bot] in #44665
fix(lens): make tool steps and conversations readable by @moe-berri in #44645
fix(ui): send null instead of $0 when a team's member default budget is cleared by @ryan-crabbe-berri in #44644
test(integration): azure_ai-route basic translation cases for six Claude models on messages, chat completions and responses by @devin-ai-integration[bot] in #44658
test(integration): basic translation cases for the openai_responses route by @devin-ai-integration[bot] in #44663
feat(lens): analyze trace workspaces with confined Python and compaction by @moe-berri in #44640
fix(vertex-ai): add deprecation_date to gemini-3.1-flash-lite-image by @berriai-litellm-provider-info-sync[bot] in #44639
fix(bedrock): add priority and flex prices for Grok 4.3, 4.6, 4.7 and Kimi K3 by @berriai-litellm-provider-info-sync[bot] in #44670
test(integration): pin the Claude-tokenizer recount for streamed claude-sonnet-5 no-usage cases by @devin-ai-integration[bot] in #44643
feat(proxy): embed enterprise LiteAdmin MCP in LiteLLM images by @tin-berri in #44610
test(integration): basic translation cases for the azure route by @devin-ai-integration[bot] in #44672
test(integration): assert the Messages API health probe on Mantle Claude by @devin-ai-integration[bot] in #44612
refactor(rust): rename litellm-framing crate to litellm-framer by @devin-ai-integration[bot] in #44683
test(integration): basic translation cases for the gemini route by @devin-ai-integration[bot] in #44662
test(integration): basic translation cases for the openai route by @devin-ai-integration[bot] in #44667
refactor(rust): derive string enum serde through strum and serde_with by @devin-ai-integration[bot] in #44675
fix(spend-tracking): stop caching failed spend-log metadata lookups as confirmed misses by @devin-ai-integration[bot] in #43560
fix(proxy): let a listed team alias win over a same-named key alias in the customer model check by @devin-ai-integration[bot] in #44677
chore: move PR template to PULL_REQUEST_TEMPLATE/general.md, add rust.md by @devin-ai-integration[bot] in #44693
fix(router): retry a /v1/messages stream the provider drops before the first content chunk by @devin-ai-integration[bot] in #44276
fix(lens): persist final coverage with source diagnostics by @moe-berri in #44682
test(integration): point the scratch upgraded proxy's read replica at the scratch database by @devin-ai-integration[bot] in #44613
fix(lens): show investigation findings for agent traces by @moe-berri in #44696
fix(build): rebuild the Rust bridge when uv sync sees Rust sources change by @nate-berri in #44708
fix(lens): preserve span timestamps in investigation evidence by @moe-berri in #44702
ci: split slow unit shards and build the Rust bridge once per run by @devin-ai-integration[bot] in #44622
test(rust_bridge): expect PartRow start_time and end_time as required LENS_CONTENT fields by @devin-ai-integration[bot] in #44719
chore(deps): bump langgraph-sdk from 0.4.2 to 0.4.4 by @dependabot[bot] in #44713
feat: add Reka as an OpenAI-compatible provider by @anniejeng in #44278
feat(ui): support native decisions endpoint in decision playground by @moyai-berriai[bot] in #44664
fix(lens): reconstruct native coding agent conversations by @moe-berri in #44711
fix(responses): honor request cache controls on chat completions bridged to the Responses API by @devin-ai-integration[bot] in #44676
fix(deps): bump source-map-js, smol-toml, mako, multidict and werkzeug for OSV advisories by @devin-ai-integration[bot] in #44728
test(integration): rename translation runner run to assert_translation by @devin-ai-integration[bot] in #44741
feat(lens): show investigation names in findings table by @moe-berri in #44753
feat(ui): configure cache-aware auto routing by @tin-berri in #43396
fix(lens): bound result recovery and preserve partial results by @moe-berri in #44692
feat(prometheus): cap series per metric for every labeled metric by @devin-ai-integration[bot] in #44420
feat(lens): type to filter the traces agent dropdown by @ishaan-berri in #44745
test(lens): use current worker protocol in billing integration by @moe-berri in #44760
feat(model_prices): add chatgpt subscription rows for the gpt-6 family by @devin-ai-integration[bot] in #44758
test(e2e): assert the sibling-replica cooldown through the router by @devin-ai-integration[bot] in #44706
test(integration): basic translation cases for the bedrock_invoke route by @devin-ai-integration[bot] in #44752
fix(gemini): stop replaying thinking block signatures to Gemini by @devin-ai-integration[bot] in #44661
fix(ci): namespace claude session ids in tracing seeds and allowlist /v1/logs on backend by @devin-ai-integration[bot] in #44761
fix(proxy): judge the free-model budget waiver by the group an alias routes to by @devin-ai-integration[bot] in #44638
test(integration): basic translation cases for the vertex_ai route by @devin-ai-integration[bot] in #44751
fix(vertex_ai): apply regional endpoint uplift on image generation cost path by @nate-berri in #44679
refactor(tests): extract Rust cache test split from #44714 by @devin-ai-integration[bot] in #44780
feat(proxy): cap batch file records, daily batch uploads, and per-file downloads by @devin-ai-integration[bot] in #43632
feat(lens): add copy link button to trace header by @ishaan-berri in #44749
test(integration): keep scripted upstream connections open past the proxy's keepalive by @devin-ai-integration[bot] in #44695
test(integration): basic translation cases for the bedrock_mantle route by @devin-ai-integration[bot] in #44750
fix(chatgpt,github_copilot): refuse device-code login inside an event loop or worker thread by @devin-ai-integration[bot] in #39585
feat(bedrock): add glm 5.3 cross-region rows and nova 2.5 sonic by @berriai-litellm-provider-info-sync[bot] in #44710
fix(responses): honor caller stream flag when provider forces SSE (internal copy of #34095) by @devin-ai-integration[bot] in #41235
refactor(cache): remove dead Cache._native_cache runtime path by @devin-ai-integration[bot] in #44785
fix(vertex_ai): add regional endpoint uplift to gemini-3.1-flash-image by @nate-berri in #44673
feat(ui): lead the Lens investigation detail with a run report by @devin-ai-integration[bot] in #44786
fix(ui): polish Lens runs loading, reload, and time range menu by @devin-ai-integration[bot] in #44789
refactor(types): move SpanAttributes, SpecialHeaders and AllowedModelRegion out of proxy._types by @devin-ai-integration[bot] in #44717
refactor(types): import proxy-only types under TYPE_CHECKING in SDK modules by @devin-ai-integration[bot] in #44740
refactor(logging): load enterprise alerting loggers lazily so import litellm skips proxy types by @devin-ai-integration[bot] in #44762
perf(types): defer pydantic schema builds via shared LiteLLMBaseModel by @devin-ai-integration[bot] in #44720
fix(router): price a model group from the deployments that serve it by @devin-ai-integration[bot] in #44732
refactor: remove fresh tech debt from the 2026-10-05 window by @devin-ai-integration[bot] in #44821
fix(azure): add gpt-6-sol priority processing prices by @berriai-litellm-provider-info-sync[bot] in #44829
refactor(types): replace Any with proven types in 6 files by @devin-ai-integration[bot] in #44491
test: deflake the JEV classifier select and two router tests that inherited leaked state by @devin-ai-integration[bot] in #44840
refactor(rust): rename litellm-core to litellm-inference by @devin-ai-integration[bot] in #44802
refactor(rust): extract inference-transcription crate by @devin-ai-integration[bot] in #44809
refactor(rust): extract inference-responses crate by @devin-ai-integration[bot] in #44811
refactor(rust): extract inference-messages crate by @devin-ai-integration[bot] in #44818
refactor(rust): extract inference-chat crate by @devin-ai-integration[bot] in #44827
refactor(rust): extract inference-ocr crate by @devin-ai-integration[bot] in #44832
chore(rust): prune inference deps and rewrite layering docs by @devin-ai-integration[bot] in #44836
fix(otel): name an Arize project on every OTel v2 Arize export by @devin-ai-integration[bot] in #44605
fix(otel): gate the Arize OTel v2 exporter on operator credentials by @devin-ai-integration[bot] in #44596
fix(otel): honor per-team Arize sampling rates in OTel v2 fan-out by @devin-ai-integration[bot] in #44595
fix(model_prices): gemini deep research input limits, vertex flash retirement dates, azure data zone gpt-6.1-sol pricing by @devin-ai-integration[bot] in #44633
refactor(types): replace Any with proven types in 157 files by @mateo-berri in #44798
fix(sap): correctly handle cache_control by @yamaceay in #39122
fix(responses): stop SDK retries nesting under router retries, and unbreak CircleCI integration tests by @devin-ai-integration[bot] in #44791
fix(cost-map): update together_ai Kimi-K3 and Qwen3.8-Flash prices to published rates by @berriai-litellm-provider-info-sync[bot] in #44864
fix(ui): restore key activity search to the top of the tab and add model activity search by @devin-ai-integration[bot] in #44521
refactor(rust): extract inference-testing crate by @devin-ai-integration[bot] in #44873
fix(ui): read MCP submission rules from bare-array /config/list response by @devin-ai-integration[bot] in #44648
chore(model_prices): remove malformed, duplicate and decommissioned palm cost map entries by @devin-ai-integration[bot] in #44880
fix(azure): add model router flat fee to azure provider cost tracking by @devin-ai-integration[bot] in #44876
chore(pricing): add azure model-router and whisper rows by @berriai-litellm-provider-info-sync[bot] in #44872
fix(security): remove the publicly known master key from the repo by @devin-ai-integration[bot] in #44718
chore(prices): add gemini/gemini-nano-banana-2.1 from the Gemini pricing page by @berriai-litellm-provider-info-sync[bot] in #44868
fix(ci): skip the generated dashboard bundle in the master key guard by @devin-ai-integration[bot] in #44890
fix(model_prices): correct supported_endpoints on gemini and vertex image rows by @devin-ai-integration[bot] in #44891
feat(vertex-ai): add gemini-nano-banana-2.1 and Nano Banana 2 priority pricing by @berriai-litellm-provider-info-sync[bot] in #44892
fix(lens): reuse trace reviews and consolidate findings across runs by @moe-berri in #44778
fix(lens): paginate visible conversation entries by @moe-berri in #44878
refactor(rust_bridge): remove rule-gated native secret-manager selection by @devin-ai-integration[bot] in #44906
feat(otel): let team and key Arize callbacks choose the OTLP transport by @devin-ai-integration[bot] in #44492
feat(proxy): add opt-in vector_store_deny_by_default for least-privilege vector store access by @devin-ai-integration[bot] in #44244
feat(auth): deny search tools by default when search_tool_deny_by_default is set by @devin-ai-integration[bot] in #44490
fix(lens): correlate native gateway spend with exact call evidence by @ishaan-berri in #44738
fix(caching): never store or serve a chat completion with no choices by @devin-ai-integration[bot] in #44709
fix(auto-router): separate tuning and select chained heuristic by @tin-berri in #44928
fix(lens): copy supported trace and span content requests by @moe-berri in #44902
fix(tracing): retain native logs and headless tool results by @moe-berri in #44894
fix(lens): preserve chronological order when grouping trace steps by @moe-berri in #44895
fix(tracing): preserve optional provider evidence in fixture replay by @moe-berri in #44930
fix(lens): use async-timeout on python 3.10 for budget reservation timeouts by @devin-ai-integration[bot] in #44911
perf(lens): reduce rendering on live investigations by @moe-berri in #44935
fix(logging): log spend once for large non-streaming requests on the chat to Responses bridge by @hektorjg in #44508
feat(azure): add azure_ai/grok-4.7 and update grok-4.6 input price by @berriai-litellm-provider-info-sync[bot] in #44937
fix(projects): show Projects to team and org admins and scope it correctly by @devin-ai-integration[bot] in #41325
fix(model_prices): drop /v1/batch from gemini/gemini-nano-banana-2.1 by @devin-ai-integration[bot] in #44940
chore: bump litellm-enterprise 0.1.73 -> 0.1.74, litellm-proxy-extras 0.4.105 -> 0.4.106, litellm 1.105.0 -> 1.106.0 by @yuneng-berri in #44942
fix(cost): price batch image output tokens at the batch image rate by @devin-ai-integration[bot] in #44897
fix(lens): refresh open traces without claiming session completion by @moe-berri in #44900
fix(proxy): bound daily spend rollup row-lock waits with lock_timeout and requeue 55P03 by @devin-ai-integration[bot] in #44450
feat(otel): trace auto-router configuration and classifier failures by @tin-berri in #44926
feat(lens): add Copy for agent to investigation details by @moyai-berriai[bot] in #44945
fix(proxy): resolve oidc/ pass-through credentials on every request by @nilpntr in #44577
feat(lens): add datasets built from real traces by @ishaan-berri in #44765
feat(lens-ui): replace the conversation view with a thread view by @ishaan-berri in #44947
fix(cli): show full-session auto-router cost comparison by @tin-berri in #44959
fix(lens): show the first user message as the run input by @ishaan-berri in #44958
fix(terraform): keep unconfigured allowed_routes plan-known and unsent by @yucheng-berri in #44487
test(e2e): add enum values, auto-discovering label gates and secret hiding for e2e metadata by @ryan-crabbe-berri in #44949
fix(proxy): stop logging license values during verification by @moyai-berriai[bot] in #44956
refactor: expose core private helpers under public names by @devin-ai-integration[bot] in #44871
test: fix shared-provider discovery, Codex catalog size and generated master key mismatches in CircleCI suites by @devin-ai-integration[bot] in #44905
chore(cost-map): sync openrouter prices from the models API by @berriai-litellm-provider-info-sync[bot] in #44975
feat(mistral): add Mistral Large 4 (Le Chonk) support by @moyai-berriai[bot] in #44870
fix(responses): preserve prompt cache reuse in chat bridge by @sorgfresser in #42281
fix(mcp): keep worker MCP configurations consistent via a catalog revision by @joshua-berri in #42568
fix(ui): give Lens traces a flush toolbar layout by @devin-ai-integration[bot] in #44972
fix(ci): generate a master key for the migration startup jobs by @devin-ai-integration[bot] in #44978
feat(mcp): add portable catalog pagination by @joshua-berri in #44446
fix(exceptions): map upstream 402 to PaymentRequiredError and cool down 402 deployments by @devin-ai-integration[bot] in #44879
feat(claude_code_gateway): issue rotating refresh tokens and a revocation endpoint by @devin-ai-integration[bot] in #44635
fix(azure): set kimi-k2.7-code deprecation date from the models list API by @berriai-litellm-provider-info-sync[bot] in #44986
fix(docker): pin openssl-3.6-dev in the pgbouncer-builder stage by @devin-ai-integration[bot] in #44991
feat(bedrock): add kimi k3 india cross-region row by @berriai-litellm-provider-info-sync[bot] in #44990
fix(azure): sync azure openai audio, realtime and image prices with azure pricing page by @berriai-litellm-provider-info-sync[bot] in #44992
test: move whole-unit legacy test files into tests/unit and delete dead skips by @devin-ai-integration[bot] in #44807
fix(bedrock): forward each Nova Sonic assistant sentence once over the realtime API by @devin-ai-integration[bot] in #44987
test(mcp): run the static root path issuer discovery test in-process by @devin-ai-integration[bot] in #44988
fix(realtime): skip guardrail VAD session.update injection for transcription sessions by @devin-ai-integration[bot] in #44843
fix(ci): stop deferred pydantic builds leaking caller locals and add the missing Lens FK migration by @devin-ai-integration[bot] in #44981
fix(ui): render team_metadata_schema keys as fixed labels by @devin-ai-integration[bot] in #41482
fix(vertex-ai): correct input_cost_per_second on gemini-3.5 transcribe-live and live-translate by @berriai-litellm-provider-info-sync[bot] in #45002
revert(docker): unpin openssl-3.6-dev in the pgbouncer-builder stage by @devin-ai-integration[bot] in #44997
test(mcp): set server_id on openapi local fake tools by @devin-ai-integration[bot] in #44998
ci: run integration-mcp on an xlarge machine by @devin-ai-integration[bot] in #44608
refactor: expose public hidden_params accessors by @devin-ai-integration[bot] in #44668
fix(azure): date claude-opus-5-5 and sonnet-5-5 and fill azure/eu/gpt-6-astra limits by @berriai-litellm-provider-info-sync[bot] in #45018
refactor: remove fresh tech debt from the 2026-10-06 window by @devin-ai-integration[bot] in #45017
ci: give the database-backed CircleCI jobs their own Postgres by @devin-ai-integration[bot] in #45007
test(proxy): make the stagger-offset and linear-dedup guards independent of runner identity and load by @devin-ai-integration[bot] in #45031
refactor(types): replace Any with proven types in 16 files by @devin-ai-integration[bot] in #45029
fix(auth): clear the recent-miss user memo when /user/new creates the user by @devin-ai-integration[bot] in #45020
fix(health): attribute background health check results to their own deployment by @devin-ai-integration[bot] in #44982
fix(ollama): turn streamed prompt-based JSON tool calls into real tool calls by @devin-ai-integration[bot] in #45053
fix(bedrock): update GovCloud OpenAI prices, add GPT-6 Astra ultrafast tier and Titan Image v2 EOL by @berriai-litellm-provider-info-sync[bot] in #45077
fix(cost-map): add us data residency multiplier to anthropic claude-opus-5-5 by @berriai-litellm-provider-info-sync[bot] in #45078
refactor(llms): expose public names for private provider helpers by @devin-ai-integration[bot] in #45037
fix(bedrock): add 2027-03-30 deprecation date to DeepSeek R1 and Qwen3 Coder rows by @berriai-litellm-provider-info-sync[bot] in #45089
fix(proxy): opt-in Redis hash-tag grouping for v3 rate limiter by @devin-ai-integration[bot] in #45085
fix(bedrock): take context and output limits from the Bedrock model cards by @berriai-litellm-provider-info-sync[bot] in #45091
fix(tests): match the lowercased bind error in the owned-proxy port-race retry by @devin-ai-integration[bot] in #45097
test(e2e): tag llm_translation tests with Subject metadata and record harness steps by @ryan-crabbe-berri in #44950
test(e2e): tag claude_code cells with Subject metadata and record CLI driver steps by @ryan-crabbe-berri in #44961
test(e2e): tag management tests with Subject metadata and record management client steps by @ryan-crabbe-berri in #44962
test(e2e): tag guardrails and logging tests with Subject metadata and record client steps by @ryan-crabbe-berri in #44963
fix(caching): skip the cache past max_messages and keep tool_result text in semantic prompts by @devin-ai-integration[bot] in #43878
test(e2e): tag router, batches and mcp tests with Subject metadata and record client steps by @ryan-crabbe-berri in #44964
test(e2e): tag the remaining quota_management tests and record budget and spend client steps by @ryan-crabbe-berri in #44966
fix(router): preserve native baseline identity and accounting by @tin-berri in #44960
test(integration): scope the team-scoped models upstream check to its own model by @devin-ai-integration[bot] in #45106
fix(prometheus): add model_group label to end-to-end latency metrics by @ahamedshaik16 in #44860
test(mcp): fix stale bridge-hook, applied-guardrails and pagination-revoke integration tests by @devin-ai-integration[bot] in #45008
test(observability): wait for warm-up spend rows in capped and count only the post-wipe half in X4 by @devin-ai-integration[bot] in #45006
fix(cost-map): lower anthropic claude-sonnet-5-5 cache read price by @berriai-litellm-provider-info-sync[bot] in #45113
fix(prompt-caching): default injected /v1/messages breakpoints to implicit lookup by @ryan-crabbe-berri in #45121
test(integration): turn the model-info refresh off in every integration proxy config by @devin-ai-integration[bot] in #45104
test(e2e): let the realtime send step accept input audio buffer frames by @ryan-crabbe-berri in #45127
feat(model_prices): add claude-haiku-5-5 model pricing by @krrish-berri-2 in #45108
fix(realtime): run transcript guardrails on raw-path transcription sessions with a transcription-safe block by @devin-ai-integration[bot] in #44844
test(integration): hold the wire barrier until the test releases it or the wire closes by @devin-ai-integration[bot] in #45124
fix(vertex-ai): correct claude-haiku-5-5 thinking and forced tool flags by @berriai-litellm-provider-info-sync[bot] in #45116
feat(helm): allow custom labels, annotations, command and args on migrationJob by @ankitasahni511 in #42242
test(e2e): tag a2a, access_control, other, secret_manager and migrations tests with Subject metadata by @ryan-crabbe-berri in #44965
fix(azure): correct azure_ai/claude-haiku-5-5 forced tool use and thinking flags by @berriai-litellm-provider-info-sync[bot] in #45114
perf(lens): claim worker jobs from an indexed due queue instead of scanning every lens by @devin-ai-integration[bot] in #45095
perf(lens): prune ClickHouse partitions when sampling and sample in one pass by @devin-ai-integration[bot] in #45087
fix(logging): bound data URI regex so base64 truncation stays linear by @devin-ai-integration[bot] in #45132
perf(lens): bound single trace reads by the sampled start time by @devin-ai-integration[bot] in #45088
feat(lens-ui): show findings ranked by priority with frequency and highlighted evidence by @ishaan-berri in #45143
fix(ollama): answer in text after a tool result and cover ollama e2e on all three endpoints by @devin-ai-integration[bot] in #45079
fix(openai-compat): send provider attribution headers on the default SDK path (+ Perplexity) by @ryan-crabbe-berri in #44291
ci(codeql): run the default suite on full scans and security-extended on PRs by @devin-ai-integration[bot] in #45149
feat(lens): flag traces with global System 1 signals by @devin-ai-integration[bot] in #45094
feat(ui): configure Anthropic workload identity federation from the dashboard by @devin-ai-integration[bot] in #44889
feat(guardrails): add logging_only_scope to observe input, output, or both by @devin-ai-integration[bot] in #43695
test: move the unit half of 126 mixed legacy files into tests/unit by @devin-ai-integration[bot] in #45090
fix(model_prices): consolidate claude-haiku-5-5 over-100k pricing and capability flags by @devin-ai-integration[bot] in #45151
feat(ui): add auto-router usage and savings table by @tin-berri in #45152
refactor(rust-bridge): unify native call inputs and Messages settings by @yujonglee-berri in #45126
test(e2e): move the harness self-tests out of tests/e2e by @ryan-crabbe-berri in #45172
test(e2e): record steps for raw transport calls and poll helpers by @ryan-crabbe-berri in #45150
feat(lens): link traces to the conversation that started them by @ishaan-berri in #45169
fix(router): resume sync streaming fallbacks without retrying primary by @amarrtech in #43959
feat(guardrails): extend Akto guardrail to responses, MCP tools, attachments and masking by @rohan-akto in #44343
fix(streaming): stop re-wrapping a bridged stream's MidStreamFallbackError by @devin-ai-integration[bot] in #44989
fix(proxy): evict cached user on every proxy for tpm/rpm updates and edit limits in the users UI by @devin-ai-integration[bot] in #44130
perf(lens): classify signals within seconds of a trace finishing by @ishaan-berri in #45186
feat(proxy): add denied_passthrough_routes deny list for custom pass-through endpoints by @devin-ai-integration[bot] in #44924
fix(responses): keep prompt_cache_breakpoint markers in the chat to responses bridge by @devin-ai-integration[bot] in #44119
fix(mcp): preserve client application type during registration by @joshua-berri in #45159
feat(mcp): rate limit all MCP operations and add server-level rpm by @devin-ai-integration[bot] in #44600
fix(ci): align misc unit tests with NativeCall bridge and widened e2e diff gates by @devin-ai-integration[bot] in #45180
fix(types): serialize deferred pydantic schema builds across threads by @devin-ai-integration[bot] in #45034
fix(anthropic): normalize images for provider token counting by @tin-berri in #45185
feat(lens): scope traces to one agent with a header picker by @ishaan-berri in #45202
chore(codeowners): replace kerry-berri with kerrylu-berri by @devin-ai-integration[bot] in #45173
test: delete 73 legacy tests covered by e2e, unable to fail, or dead in CI by @devin-ai-integration[bot] in #45195
feat(lens): isolate ingestion and investigations in a Rust service by @moe-berri in #45148
fix(mcp): honor scoped cache freshness by @joshua-berri in #45165
chore(ui): bump next to 16.3.8 by @yuneng-berri in #45207
refactor(rust-bridge): share field and response marshaling by @yujonglee-berri in #45187
fix(lens): preserve numeric tags and report OTLP error codes by @moe-berri in #45206
fix(scheduler): remove a request's queue entry once it stops waiting by @RachelHuangZW in #43061
fix(vertex_ai): dial the multi-region Live API host for realtime sessions by @mateo-berri in #45166
feat(proxy,ui): add Moyai to the view switcher with quick connect by @ishaan-berri in #45196
fix(ci): restore vertex model sets mutated by get_optional_params tests by @devin-ai-integration[bot] in #45209
feat(decisions): add the OpenAI Decisions spec types and the System One translation by @devin-ai-integration[bot] in #45129
fix(anthropic): let /v1/messages mid-stream failures reach the proxy failure boundary by @Winoren in #44800
test: remove dead imports and helpers left behind by legacy test deletion by @devin-ai-integration[bot] in #45208
fix(vertex-ai): halve claude-sonnet-5-5 cache read price by @berriai-litellm-provider-info-sync[bot] in #45217
fix(anthropic): fail closed when a token-file or inline-token federation credential is missing a rule or organization id by @mateo-berri in #45178
fix(bedrock): map Scheduled batch jobs to in_progress on retrieve by @mrinal-berri in #45157
feat(ui): redesign usage page with stacked model and agent charts by @ishaan-berri in #45221
fix(bedrock): lower claude sonnet 5.5 cache read price to the aws offer index by @berriai-litellm-provider-info-sync[bot] in #45216
fix(responses): keep cache breakpoints on blocks the bridge stringifies by @mateo-berri in #40032
fix(bedrock_mantle): lower claude sonnet 5.5 cache read price to match bedrock runtime by @berriai-litellm-provider-info-sync[bot] in #45223
fix(ui): address usage redesign review findings by @ishaan-berri in #45229
perf(lens): faster trace opens and list pages at scale by @ishaan-berri in #45228
refactor(proxy): expose public names for private proxy helpers by @mateo-berri in #45170
fix(ollama): one stream id, 400 on non-text tool content, separated tool results, no empty message item by @mateo-berri in #45177
feat(lens): simplify deployment and first trace setup by @moe-berri in #45230
fix(proxy): traceparent/baggage fallback must not override caller metadata by @devin-ai-integration[bot] in #43688
fix(proxy): accept router-wide default_litellm_params in required-param validation by @yucheng-berri in #44483
fix(proxy): keep idle responses websockets open until a configurable session limit by @devin-ai-integration[bot] in #44433
fix(lens-ui): open traces on the top-level step and stop showing model names as input by @ishaan-berri in #45254
refactor(decisions): dispatch /v1/decisions through provider configs and the shared HTTP handler by @devin-ai-integration[bot] in #45130
feat(lens): show agent, user and slack thread first in the run header by @ishaan-berri in #45261
fix(proxy): declare the Moyai settings write's service target and allowlist its routes by @devin-ai-integration[bot] in #45220
feat(lens): show end-user feedback on traces, stored in ClickHouse by @ishaan-berri in #45171
fix(proxy): admit litellm_proxy/hosted_vllm in provider-endpoint discovery by @mayuriphad in #38617
feat(decisions): serve System One format at /v1/systemone and OpenAI format at /v1/decisions by @mateo-berri in #45184
feat(decisions): add OpenAI as a Decisions provider behind a shared decisions format by @mateo-berri in #45214
feat(lint): add LIT015 requiring pydantic models to be frozen by @devin-ai-integration[bot] in #42348
fix(router): keep include_fallback_errors off the provider call on the sync Router path by @mateo-berri in #45218
fix(router): honor a per-request fallbacks list on a mid-stream fallback by @mateo-berri in #45227
fix(responses): map input_audio blocks in the chat-to-Responses bridge by @mateo-berri in #45224
fix(proxy): answer a rejected websocket handshake without crashing the HTTP exception handler by @mateo-berri in #45251
ci(lint): gate every rule on its merge-base count and cap Anys at a fixed total by @devin-ai-integration[bot] in #40193
test(e2e): cover responses API gaps on gemini, azure, compact and context management by @devin-ai-integration[bot] in #45248
chore: bump litellm-proxy-extras to 0.4.107 and litellm-enterprise to 0.1.75 by @devin-ai-integration[bot] in #45291
test(integration): boot owned proxies with a readiness-sized worker healthcheck budget by @devin-ai-integration[bot] in #45292
fix(utils): make supports_audio_output read the supports_audio_output cost-map key by @devin-ai-integration[bot] in #45294
fix(lens): refresh runs until gateway costs are complete by @tin-berri in #45105
fix(lint-gates): fail on an empty base scan instead of blaming the change for every violation by @devin-ai-integration[bot] in #45314
refactor(proxy): type fresh Moyai connect helpers and drop MCP rpm getattr by @devin-ai-integration[bot] in #45315
revert(lint-gates): accept an empty base scan again, since zero violations is a legitimate count by @devin-ai-integration[bot] in #45319
fix(bedrock_mantle): send OpenAI explicit prompt cache breakpoints for GPT-5.6 and newer by @msdianprince-7 in #38729
test: make 77 legacy live tests offline in litellm_utils, router_unit and responses dirs by @devin-ai-integration[bot] in #45298
test: move 81 legacy live tests in pass-through, spend, batches, audio, search, guardrails, image and ocr dirs offline by @devin-ai-integration[bot] in #45288
feat(ui): add TypeSafe and Strands Decider to the Add Model provider list by @devin-ai-integration[bot] in #45335
feat(mcp): support upstream OAuth client metadata identities by @joshua-berri in #45231
fix(proxy-extras): name the database error when the Lens rename check cannot run by @devin-ai-integration[bot] in #45341
test(integration): run the response cache tool-call tests on a proxy without the message cap by @devin-ai-integration[bot] in #45350
fix(logging): skip sync success callbacks for internal sub-calls, deflake RAG and Langfuse tests by @devin-ai-integration[bot] in #45345
fix(azure): move sora-2 retirement date to the later Models API date by @berriai-litellm-provider-info-sync[bot] in #45355
fix(lens): block teamless feedback writes and fix retention test by @ishaan-berri in #45266
fix(ci): trim rust-test debuginfo so the job fits on the runner disk by @devin-ai-integration[bot] in #45318
fix(ui): list only catalog models in the Add Model picker by @devin-ai-integration[bot] in #44877
feat(rust): add standalone typed LLM wire contracts by @yujonglee-berri in #45188
test: delete 34 legacy tests already covered by e2e by @devin-ai-integration[bot] in #45340
fix(responses): keep tool_result next to tool_use on Anthropic previous_response_id continuations by @mateo-berri in #45322
test: realign stale decisions and lens tests with merged behaviour by @devin-ai-integration[bot] in #45346
feat(prometheus): expose per-project per-model rate limit allowed and used gauges by @devin-ai-integration[bot] in #43561
fix(mcp): preserve elicitation context and report relay failures by @joshua-berri in #45255
docs(rust): add ADRs for the Rust core by @nate-berri in #45205
test: move 76 live top-level tests to offline unit and integration coverage by @devin-ai-integration[bot] in #45352
fix(proxy): mark background responses stale_expired when provider returns 404 by @devin-ai-integration[bot] in #41724
fix(bedrock): take context, output limits and EOL dates from the Bedrock model cards by @berriai-litellm-provider-info-sync[bot] in #45409
chore(cost-map): add openai gpt-6.1-sol ultrafast tier prices from the pricing page by @berriai-litellm-provider-info-sync[bot] in #45408
fix(bedrock_mantle): take gpt-oss output limits and EOL dates from the Bedrock model cards by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45417
fix(rust): path-qualify attribute aliases so rust-analyzer resolves them by @nate-berri in https://github.com/BerriAI/litellm/pull/45416
fix(proxy): use budget reset window for projected spend alerts by @Solomon-mithra in https://github.com/BerriAI/litellm/pull/31942
test(proxy): audit the websocket rejection log on every live route by @mateo-berri in https://github.com/BerriAI/litellm/pull/45293
fix(proxy): retry lock-timed-out daily spend batches in place so the shutdown flush keeps them by @mateo-berri in https://github.com/BerriAI/litellm/pull/45273
fix(vertex_ai): forward the inline-tools-2026-09-15 beta to Vertex and Anthropic by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45297
feat(fireworks_ai): forward the LiteLLM user id as user behind fireworks_forward_user_id by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45265
fix(model_prices): add Bedrock flex prices, Anthropic web search flags, Gemini shutdown dates, OpenRouter alias drift, Mistral Large 4 context by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/44910
feat(voyage): rebrand to VoyageAI by MongoDB and route MongoDB keys to ai.mongodb.com by @fzowl in https://github.com/BerriAI/litellm/pull/41812
feat(gemini): accept file content blocks with video_metadata on multimodal embeddings by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45305
feat(rust): add the Anthropic beta header policy by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/45419
test(decisions): post System One bodies to /v1/systemone in the translation bases by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45327
fix(ci): drop the publicly known master key prefix from the voyage routing test by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45431
feat(guardrails): per-mode stream_scope with bedrock stream and pass-through fixes by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/43801
refactor(python-bridge): take NativeCall directly and fold routes into per-route folders by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/45413
fix(lens): prevent progress updates from starving analysis budget reservations by @moe-berri in https://github.com/BerriAI/litellm/pull/45432
refactor(rust): derive strum VariantArray and string conversions by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/45434
chore(decisions): remove the System One converters and OpenAI spec types left dead by #45214 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45442
feat(spend_logs): configure which metadata fields are stored in LiteLLM_SpendLogs by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/44659
fix(ui): show the Add Model picker once the model catalog loads after a provider is picked by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45426
fix(rust): add the inline-tools-2026-09-15 beta to AnthropicBeta by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45439
fix(bedrock): count tokens on bedrock-mantle when bedrock-runtime cannot count a Claude model by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45317
fix(ui): link model access group chips to the access group filter by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45402
fix(proxy): accept team-scoped models by their public name on POST /fallback by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45455
fix(otel): emit OpenInference tool calls and metadata on Arize OTel v2 spans by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/43698
test(ui-e2e): update usage page selectors after the #45221 redesign by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45347
test: replace 61 live logging and otel tests with offline unit and integration coverage by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45362
fix(bedrock): remove the bare xai.grok-4.7 cost-map row that AWS cannot invoke by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45460
test: make 38 legacy live base-class translation tests offline by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45351
fix(bedrock): split the tag for gpt-oss only on native Chat Completions by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/44432
feat(ui): add Cmd+K command palette with key search by @ishaan-berri in https://github.com/BerriAI/litellm/pull/45456
test(streaming): expect the unwrapped provider error in bridged /v1/messages error frames by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45465
refactor(python-bridge): derive NativeCall extraction by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/45449
refactor(python-bridge): ship a signature base and read the resolved call by @yujonglee-berri in https://github.com/BerriAI/litellm/pull/45450
fix(cost-map): update baseten DeepSeek-V4.1-Flash-Fast cache read price and max output by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45468
ci(unit): add unit passed collector job and fold proxy-db shards into test-unit.yml by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45466
feat(guardrails): support logging_only mode for the Akto guardrail by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45461
fix(router): drop the encrypted reasoning a fallback hop's target cannot decrypt by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45393
chore(cost-map): sync openrouter prices from the models API by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45469
fix(cli): pin pi compat flags so lite pi stops sending store to Anthropic models by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/40739
feat(bedrock): add twelvelabs pegasus 1.5 inference profiles by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45477
feat(sdk): build core from an independent packaging manifest by @joshua-berri in https://github.com/BerriAI/litellm/pull/44340
fix(proxy): save file details for every batch output file so they list and retrieve by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/41761
fix(bedrock): remove the bare openai.gpt-6.1-sol cost-map row that AWS cannot invoke by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45479
refactor(sdk): separate core AWS and tokenizer dependencies by @joshua-berri in https://github.com/BerriAI/litellm/pull/44447
fix(router): match deployment pricing ids against the cost map only within the deployment's provider by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45472
fix(bedrock): add gpt-6.1-sol ultrafast tier prices from the Bedrock model card by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45482
fix(ci): import seed_tracing_fixtures from the pytest scripts path in rust trace tests by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45485
fix(bedrock): take gpt-6.1-sol context window from the Bedrock model card by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45488
feat(ui): add the evaluation mode to the Add Model form by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45481
feat(mcp): translate input requests and bind continuations by @joshua-berri in https://github.com/BerriAI/litellm/pull/45464
fix(bedrock): add the claude-sonnet-4-5 EOL date from the Bedrock model card by @berriai-litellm-provider-info-sync[bot] in https://github.com/BerriAI/litellm/pull/45496
fix(token_counter): price a base64 PDF document per page instead of as one image by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45301
ci: render lint, unit and smoke checks as / with one collector per tier by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45480
test(routing): wait on deployment registration instead of boot model-info traffic by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45267
fix(anthropic): count a leading system run through count_tokens' system parameter by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45463
ci: run only integration jobs on PR CircleCI pipelines and move router and guardrails suites to GHA by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45509
test: update stale MCP call_tool and usage card assertions, add timeout headroom to request-log index boot test by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45523
feat(credentials): add display_name and make credential_name immutable on PATCH by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/43148
fix(ui): let the model picker remove selections that are no longer available by @yuneng-berri in https://github.com/BerriAI/litellm/pull/43944
refactor(proxy): inject one UTC clock read per operation into gateway tracking, PTU rollup and Mavvrik export by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45520
refactor(proxy): inject the clock into the v1 parallel request limiter and pin it in its tests by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45522
chore: bump litellm-enterprise 0.1.75 -> 0.1.76, litellm-proxy-extras 0.4.107 -> 0.4.108 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45534
feat(ui): configure OpenAI workload identity federation from the LLM Credentials and Add Model forms by @mateo-berri in https://github.com/BerriAI/litellm/pull/45528
fix(mcp): never forward the caller's LiteLLM key to MCP servers by @joshua-berri in https://github.com/BerriAI/litellm/pull/45401
ci: run every build_and_test job in every CircleCI pipeline (#43152) by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45651
feat(decisions): backport Microsoft-Decision-1 and the decision models Admin UI to rc/1.106.0 (#45639, #45673, #45654, #45717) by @mateo-berri in https://github.com/BerriAI/litellm/pull/45735
feat(ui): backport the System One request form to rc/1.106.0 (#45721) by @mateo-berri in https://github.com/BerriAI/litellm/pull/45739
feat(guardrails): backport the logging_only scope split to rc/1.106.0 (#45709) by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45782
fix(ui): backport modal backdrop fix to rc/1.106.0 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45877
feat(guardrails): backport the decision model guardrail to rc/1.106.0 (#45663) by @yucheng-berri in https://github.com/BerriAI/litellm/pull/45888
fix(credentials): backport credential edit fixes to rc/1.106.0 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/45890
fix(mcp): backport request-local guardrail bearers to RC 106 by @joshua-berri in https://github.com/BerriAI/litellm/pull/45901
feat(decisions): backport the rest of the decisions and System One launch to rc/1.106.0 by @mateo-berri in https://github.com/BerriAI/litellm/pull/45891
feat(ui): backport the Home page dashboard root to rc/1.106.0 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45911
feat(ui): backport decision auto-router setup to 1.106 by @joshua-berri in https://github.com/BerriAI/litellm/pull/45916
fix(home): serve the bundled whats_new.json by default on rc/1.106.0 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45922
feat(ui): backport the form-first Decisions playground restyle to rc/1.106.0 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45929
chore: rebuild Admin UI bundle for rc/1.106.0 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45930
fix(ui): backport Decisions yes / no answer type label to rc/1.106.0 (#45931) by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/45933
feat(ui): backport #46260 Home What's new from day 0 support blog posts to rc/1.106.0 by @kerry-berri in https://github.com/BerriAI/litellm/pull/46309
fix(guardrails): backport #46086 to rc/1.106.0 by @mateo-berri in https://github.com/BerriAI/litellm/pull/46383
fix(prometheus): backport #46267 to rc/1.106.0 by @yucheng-berri in https://github.com/BerriAI/litellm/pull/46491
feat(decisions): backport #45474, #45811, #45734, #46594, and #46596 to rc/1.106.0 by @devin-ai-integration[bot] in https://github.com/BerriAI/litellm/pull/46628
fix(proxy): backport #46408 to rc/1.106.0 by @mateo-berri in https://github.com/BerriAI/litellm/pull/46827
fix(caching): backport #45784 and #46488 to rc/1.106.0 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/47006
test(e2e): backport three stale e2e test fixes from main to rc/1.106.0 by @yuneng-berri in https://github.com/BerriAI/litellm/pull/47229

New Contributors

@Bernedotcom2312 made their first contribution in #42141
@jimaldon made their first contribution in #40924
@SteffanWolter made their first contribution in #40108
@wakqasahmed made their first contribution in #42213
@Aasif-Multani made their first contribution in #44183
@joeym82956 made their first contribution in #44318
@kohjunhao made their first contribution in #44173
@ninadphalak made their first contribution in #42645
@anniejeng made their first contribution in #44278
@yamaceay made their first contribution in #39122
@hektorjg made their first contribution in #44508
@nilpntr made their first contribution in #44577
@sorgfresser made their first contribution in #42281
@ankitasahni511 made their first contribution in #42242
@amarrtech made their first contribution in #43959
@rohan-akto made their first contribution in #44343
@RachelHuangZW made their first contribution in #43061
@Winoren made their first contribution in #44800
@mayuriphad made their first contribution in #38617
@Solomon-mithra made their first contribution in https://github.com/BerriAI/litellm/pull/31942

Full Changelog: v1.105.0...v1.106.0