PyPI · #3877 most downloaded on PyPI

langchain-litellm

0.11.01.1M downloads/molangchain-ai/langchain-litellm ↗New releases in your daily digest, changelog included.

An integration package connecting LiteLLM and LangChain

Last release 4 days ago
30 Sep 2026
Release timing varies
gaps range from 8 days to 3 months
Nearly every release is documented
notes for 32 of 32 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
32 releases · first in 2025

Release timeline

32 releases since 2025

20252026

One column per month.

Releases

0.11.030 Sept 20262 sources agree
chat_models: a streamed reply carries response-level provider_specific_fields, including Vertex grounding metadata, in response_metadata rather than a
0.11.0 (2026-09-30)

⚠ BREAKING CHANGES

chat_models: a streamed reply carries response-level provider_specific_fields, including Vertex grounding metadata, in response_metadata rather than additional_kwargs, as invoke does. Read message.response_metadata["provider_specific_fields"] on both paths.

LiteLLMEmbeddingsRouter raises ValueError for a set api_base or base_url, which it used to accept and silently ignore. Configure the endpoint on the Router deployment instead.

Features

chat_models: carry Responses API reasoning items across turns (#318) (53bce51), closes #303

chat_models: make use_responses_api on ChatLiteLLMRouter a check (#333) (892fae6), closes #324

chat_models: surface url_citation annotations on replies (#325) (1efb90a), closes #304

Bug Fixes

chat_models: report a Router stream's real cost after a fallback (#341) (03694b2)

chat_models: stream response-level provider_specific_fields in response_metadata (#342) (f94c52e)

deps: raise the pydantic and typing-extensions floors to installable ones (#345) (9ac3bef)

deps: require litellm 1.101.0 for a streamed reply's cost (#339) (5bf9edd)

document_loaders: read null page dimensions as absent in LiteLLMOCRLoader (#332) (86bf2b5), closes #331

standardize endpoint URL handling (#320) (513a3af), closes #288

What's Changed

test(chat_models): add cache_write_tokens to the Responses API usage fixture by Akshay Dongare (@Akshay-Dongare) in #321

feat(chat_models): carry Responses API reasoning items across turns by Akshay Dongare (@Akshay-Dongare) in #318

fix!: standardize endpoint URL handling by Dinesh Jat (@dineshjakhad) in #320

build(deps): bump the github-actions group with 3 updates by Dependabot (@dependabot)[bot] in #326

build(deps): bump the python-deps group with 6 updates by Dependabot (@dependabot)[bot] in #327

feat(chat_models): surface url_citation annotations on replies by Akshay Dongare (@Akshay-Dongare) in #325

fix(document_loaders): read null page dimensions as absent in LiteLLMOCRLoader by Akshay Dongare (@Akshay-Dongare) in #332

feat(chat_models): make use_responses_api on ChatLiteLLMRouter a check by Akshay Dongare (@Akshay-Dongare) in #333

fix(deps): require litellm 1.101.0 for a streamed reply's cost by Akshay Dongare (@Akshay-Dongare) in #339

build(deps): bump urllib3 from 2.7.0 to 2.8.0 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #340

fix(chat_models)!: stream response-level provider_specific_fields in response_metadata by Akshay Dongare (@Akshay-Dongare) in #342

fix(chat_models): report a Router stream's real cost after a fallback by Akshay Dongare (@Akshay-Dongare) in #341

fix(deps): raise the pydantic and typing-extensions floors to installable ones by Akshay Dongare (@Akshay-Dongare) in #345

chore(main): release langchain-litellm 0.11.0 by @github-actions[bot] in #322

Full Changelog: langchain-litellm==0.10.0...langchain-litellm==0.11.0

Additional notes

langchain-litellm: v0.11.0 Latest

Latest

Compare

Choose a tag to compare

Open source →

0.10.027 Sept 20262 sources agree

chat_models: a tool_choice string other than auto, none, required or any is read as a tool name and must name a bound function tool, or bind_tools rai

0.10.0 (2026-09-27)

⚠ BREAKING CHANGES

chat_models: a tool_choice string other than auto, none, required or any is read as a tool name and must name a bound function tool, or bind_tools raises ValueError. On mistral/, codestral/ and Vertex AI Mistral models a tool name no longer forces a tool call, since litellm drops the function choice there; pass "required" to keep forcing. On Claude, a tool name is now forced even where a forced tool is refused. The downgrade to "auto" sees only manual thinking known when bind_tools runs, on a model it can tell is Claude, so it misses thinking or reasoning_effort passed at invoke time; a model named through litellm.model_alias_map or a Router model_group_alias; thinking that a wildcard Router deployment or a Router fallback adds; and a LiteLLM proxy that adds thinking or makes it from reasoning_effort. Opus 5.5, Fable 5.1 and Mythos 5.1 refuse a forced tool even without thinking. Pass "auto" in all of these cases.
chat_models: use_responses_api=True, which ChatLiteLLM ignored, now sends calls to the Responses API, where litellm drops Chat Completions-only params such as stop, n and seed, and raises ValueError for a model litellm cannot bridge; ChatLiteLLMRouter raises ValueError for it. Drop the flag to stay on Chat Completions.

Features

chat_models: route calls through the Responses API with use_responses_api (#301) (2a7a4a2)

Bug Fixes

chat_models: downgrade a forced tool_choice wherever thinking is manual (#313) (079aa08)
chat_models: drop normalized reasoning blocks from outbound content (#317) (496c863)
chat_models: force the tool a string tool_choice names (#311) (cad44b6)
chat_models: rejoin a reply litellm splits across choices (#307) (a6fe164)
chat_models: replay signed thinking blocks to the endpoint that signed them (#300) (5cfd986)
chat_models: stop bind_tools raising KeyError on a dict tool_choice (#310) (9db3498)

What's Changed

chore: adopt ruff 0.16's default rule set, pinned by full code by Akshay Dongare (@Akshay-Dongare) in #297
chore: install the locked ruff with a plain uv sync by Akshay Dongare (@Akshay-Dongare) in #299
fix(chat_models): rejoin a reply litellm splits across choices by Akshay Dongare (@Akshay-Dongare) in #307
feat(chat_models)!: route calls through the Responses API with use_responses_api by Akshay Dongare (@Akshay-Dongare) in #301
fix(chat_models): replay signed thinking blocks to the endpoint that signed them by Akshay Dongare (@Akshay-Dongare) in #300
docs(chat_models): show how ChatLiteLLMRouter reaches OpenAI's built-in tools by Akshay Dongare (@Akshay-Dongare) in #309
fix(chat_models): stop bind_tools raising KeyError on a dict tool_choice by Akshay Dongare (@Akshay-Dongare) in #310
fix(chat_models)!: force the tool a string tool_choice names by Akshay Dongare (@Akshay-Dongare) in #311
fix(chat_models): downgrade a forced tool_choice wherever thinking is manual by Akshay Dongare (@Akshay-Dongare) in #313
fix(chat_models): drop normalized reasoning blocks from outbound content by Feng Xiao (@iam-xiaofeng) in #317
chore(main): release langchain-litellm 0.10.0 by @github-actions[bot] in #308

New Contributors

Feng Xiao (@iam-xiaofeng) made their first contribution in #317

Full Changelog: langchain-litellm==0.9.0...langchain-litellm==0.10.0

Open source →

Additional notes

langchain-litellm: v0.10.0

Compare

Choose a tag to compare

Open source →

0.9.024 Sept 20262 sources agree

langchain-litellm now requires Python 3.11 or later. Stay on langchain-litellm 0.8.x to keep using Python 3.10.

0.9.0 (2026-09-24)

⚠ BREAKING CHANGES

langchain-litellm now requires Python 3.11 or later. Stay on langchain-litellm 0.8.x to keep using Python 3.10.

Features

drop support for Python 3.10 (3d75c58)

What's Changed

feat!: drop support for Python 3.10 by Akshay Dongare (@Akshay-Dongare) in #293
ci: test on Python 3.14 by Akshay Dongare (@Akshay-Dongare) in #296
chore(main): release langchain-litellm 0.9.0 by @github-actions[bot] in #295

Full Changelog: langchain-litellm==0.8.1...langchain-litellm==0.9.0

Open source →

Additional notes

langchain-litellm: v0.9.0

Compare

Choose a tag to compare

Open source →

0.8.124 Sept 20262 sources agree

chat_models: preserve file metadata after normalization

0.8.1 (2026-09-24)

Bug Fixes

chat_models: preserve file metadata after normalization (#289) (46abd65)
deps: exclude litellm releases that fail to import on Python 3.10 (#290) (092b466)

What's Changed

docs(changelog): reconstruct the history before 0.6.0 by Akshay Dongare (@Akshay-Dongare) in #285
ci(release): fail when a releasable commit is missing from the notes by Akshay Dongare (@Akshay-Dongare) in #286
fix(chat_models): preserve file metadata after normalization by Dinesh Jat (@dineshjakhad) in #289
fix(deps): exclude litellm releases that fail to import on Python 3.10 by Akshay Dongare (@Akshay-Dongare) in #290
chore(main): release langchain-litellm 0.8.1 by @github-actions[bot] in #291

Full Changelog: langchain-litellm==0.8.0...langchain-litellm==0.8.1

Open source →

Additional notes

langchain-litellm: v0.8.1

Compare

Choose a tag to compare

Open source →

0.8.023 Sept 20262 sources agree

chat_models: .stream() and .astream() now stream incrementally on an instance that did not pass streaming=True . They previously returned the whole re

0.8.0 (2026-09-23)

⚠ BREAKING CHANGES

chat_models: .stream() and .astream() now stream incrementally on an instance that did not pass streaming=True. They previously returned the whole response as a single chunk, because the default streaming=False was recorded as an explicit opt-out. Code that assumed one chunk per call, or that expected astream_events to emit a single on_chat_model_stream, now receives one per token. Pass streaming=False explicitly to keep the previous behaviour.
embeddings: LiteLLMEmbeddings rejects unknown constructor kwargs rather than discarding them. timeout, max_tokens, client and streaming were accepted and ignored; use the declared request_timeout, or pass provider values through model_kwargs.
chat_models: ChatLiteLLMRouter no longer copies litellm's router metadata into response_metadata. Read response_cost instead of hidden_params.response_cost, and model_id instead of hidden_params.model_id; both are now present on the streaming paths too, where none of these keys ever appeared. api_base, attempted_fallbacks, attempted_retries, caching_groups, deployment, deployment_model_name, hidden_params, max_retries, model_group, model_group_alias, model_group_size, model_info and original_model_group are gone with no replacement.

Features

chat_models: surface the response cost in response_metadata (#281) (5d04bb8)
make .stream() stream, honour embeddings retries, and stop dropping credentials (#272) (08cc7be)

Bug Fixes

accept base_url as an alias for api_base in ChatLiteLLM (#200) (14d8c07)
accept base_url for LiteLLM embeddings (#203) (e5b2e5e)
chat_models: forward provider-specific api_key fields to litellm (#261) (b57c1e4)
chat_models: honor per-call model override in _get_ls_params (#248) (c2d4fec)
chat_models: include top_p and top_k in _default_params (#233) (a03a841)
chat_models: name a streamed cost once (#284) (919944d)
chat_models: omit top_p and top_k when the caller left them unset (#278) (e0c222e)
chat_models: report unparsable tool-call arguments as invalid (#260) (c93299f)
chat_models: stop echoing an unparsable tool call back to the provider (#279) (a3bf8ec)
chat_models: stop injecting thinking blocks into AIMessage.content (#244) (a47db29)
chat_models: stop republishing litellm's router bookkeeping (#283) (c568593)
chat_models: surface finish_reason in streamed response_metadata (#241) (b0ca0c3)
deps: declare pydantic and typing-extensions (#280) (57ae70e)
router: honour max_retries in ChatLiteLLMRouter (#262) (dd8b4f1)

What's Changed

fix(chat_models): include top_p and top_k in _default_params by Ashutosh Dongare (@AshutoshDongare) in #233
test(chat_models): pin that top_p and top_k default to None by SpiliosDmk (@SpiliosDimakopoulos) in #240
fix: accept base_url for LiteLLM embeddings by Bing Gao (@binggao1230) in #203
fix: accept base_url as an alias for api_base in ChatLiteLLM by Tiago Ferreira (@tcferreira) in #200
fix(chat_models): stop injecting thinking blocks into AIMessage.content by Benjamin Silva (@bssilva06) in #244
fix(chat_models): report unparsable tool-call arguments as invalid by Shyamji Pandey (@ShyamRV) in #260
fix(router): honour max_retries in ChatLiteLLMRouter by Dinesh Jat (@dineshjakhad) in #262
build(deps): bump anyio from 4.12.1 to 4.14.2 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #273
docs(readme): point activity badges at the canonical repo by Akshay Dongare (@Akshay-Dongare) in #274
docs(pyproject): add a maintainer link to the PyPI project URLs by Akshay Dongare (@Akshay-Dongare) in #275
fix(chat_models): honor per-call model override in _get_ls_params by Shyamji Pandey (@ShyamRV) in #248
build(deps): bump the github-actions group with 2 updates by Dependabot (@dependabot)[bot] in #276
build(deps-dev): bump the python-deps group with 2 updates by Dependabot (@dependabot)[bot] in #277
feat!: make .stream() stream, honour embeddings retries, and stop dropping credentials by Akshay Dongare (@Akshay-Dongare) in #272
fix(chat_models): forward provider-specific api_key fields to litellm by Shyamji Pandey (@ShyamRV) in #261
fix(chat_models): omit top_p and top_k when the caller left them unset by Akshay Dongare (@Akshay-Dongare) in #278
fix(chat_models): stop echoing an unparsable tool call back to the provider by Akshay Dongare (@Akshay-Dongare) in #279
fix(chat_models): surface finish_reason in streamed response_metadata by SpiliosDmk (@SpiliosDimakopoulos) in #241
fix(deps): declare pydantic and typing-extensions by Akshay Dongare (@Akshay-Dongare) in #280
feat(chat_models): surface the response cost in response_metadata by Akshay Dongare (@Akshay-Dongare) in #281
fix(chat_models)!: stop republishing litellm's router bookkeeping by Akshay Dongare (@Akshay-Dongare) in #283
fix(chat_models): name a streamed cost once by Akshay Dongare (@Akshay-Dongare) in #284
chore(main): release langchain-litellm 0.8.0 by @github-actions[bot] in #270

New Contributors

SpiliosDmk (@SpiliosDimakopoulos) made their first contribution in #240
Bing Gao (@binggao1230) made their first contribution in #203
Tiago Ferreira (@tcferreira) made their first contribution in #200
Benjamin Silva (@bssilva06) made their first contribution in #244
Dinesh Jat (@dineshjakhad) made their first contribution in #262

Full Changelog: langchain-litellm==0.7.2...langchain-litellm==0.8.0

Open source →

Additional notes

langchain-litellm: v0.8.0

Compare

Choose a tag to compare

Open source →

0.7.216 Sept 2026Deprecated3 sources agree

test(scripts): cover check_imports against shadowing and the deprecated loader by Akshay Dongare (@Akshay-Dongare) in #268

0.7.2 (2026-09-16)

Bug Fixes

deps: drop the unused cryptography dependency (#267) (8aa4fc6)

What's Changed

build(deps): bump tornado from 6.5.7 to 6.5.8 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #258
chore(lint): select a minimal ruff rule set by Shyamji Pandey (@ShyamRV) in #254
ci: run lint, type and spell checks on every pull request by Akshay Dongare (@Akshay-Dongare) in #266
chore(github): move issue and PR templates into .github by Shyamji Pandey (@ShyamRV) in #253
build(deps): bump the python-deps group across 1 directory with 11 updates by Dependabot (@dependabot)[bot] in #264
build(deps): bump the github-actions group across 1 directory with 3 updates by Dependabot (@dependabot)[bot] in #237
chore(scripts): replace load_module() in check_imports, removed in Python 3.15 by Shyamji Pandey (@ShyamRV) in #259
chore(makefile): repair lint_diff and format_diff after the repo extraction by Shyamji Pandey (@ShyamRV) in #255
fix(deps): drop the unused cryptography dependency by Akshay Dongare (@Akshay-Dongare) in #267
test(scripts): cover check_imports against shadowing and the deprecated loader by Akshay Dongare (@Akshay-Dongare) in #268
chore(main): release langchain-litellm 0.7.2 by @github-actions[bot] in #269

Full Changelog: langchain-litellm==0.7.1...langchain-litellm==0.7.2

Open source →

Additional notes

langchain-litellm: v0.7.2

Compare

Choose a tag to compare

Open source →

0.7.131 Aug 20263 sources agree

ci(release): gate PyPI publish via environment instead of draft releases by Akshay Dongare (@Akshay-Dongare) in #195

0.7.1 (2026-08-31)

Bug Fixes

litellm: bump langchain-core (#245) (060fe0c)

What's Changed

ci(release): gate PyPI publish via environment instead of draft releases by Akshay Dongare (@Akshay-Dongare) in #195
build(deps): bump the python-deps group with 5 updates by Dependabot (@dependabot)[bot] in #196
build(deps): bump the python-security group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #198
chore(deps): bump langchain to 1.3.10 by Akshay Dongare (@Akshay-Dongare) in #199
ci(codeql): Add CodeQL analysis workflow by Akshay Dongare (@Akshay-Dongare) in #206
ci(codeql): switch to advanced setup to support fork PRs by Akshay Dongare (@Akshay-Dongare) in #205
build(deps): bump the python-security group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #207
build(deps): bump the python-deps group with 3 updates by Dependabot (@dependabot)[bot] in #210
build(deps): bump actions/checkout from 6.0.3 to 7.0.0 in the github-actions group by Dependabot (@dependabot)[bot] in #209
build(deps): bump langgraph-checkpoint from 4.1.0 to 4.1.1 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #211
build(deps): bump the python-deps group with 3 updates by Dependabot (@dependabot)[bot] in #214
build(deps): bump actions/setup-python from 6.2.0 to 6.3.0 in the github-actions group by Dependabot (@dependabot)[bot] in #213
build(deps): bump the github-actions group with 3 updates by Dependabot (@dependabot)[bot] in #219
build(deps): bump the python-deps group across 1 directory with 7 updates by Dependabot (@dependabot)[bot] in #221
build(deps): bump the github-actions group with 5 updates by Dependabot (@dependabot)[bot] in #227
build(deps): bump the python-deps group across 1 directory with 5 updates by Dependabot (@dependabot)[bot] in #229
build(deps): bump the python-security group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #230
fix(litellm): bump langchain-core by ccurme (@ccurme) in #245
chore(main): release langchain-litellm 0.7.1 by @github-actions[bot] in #246

New Contributors

ccurme (@ccurme) made their first contribution in #245

Full Changelog: langchain-litellm==0.7.0...langchain-litellm==0.7.1

Open source →

Additional notes

langchain-litellm: v0.7.1

Compare

Choose a tag to compare

Open source →

0.7.015 Jun 20263 sources agree

langchain-litellm: record package version in model metadata

0.7.0 (2026-06-15)

Features

langchain-litellm: record package version in model metadata (#192) (1c0dcf3)

What's Changed

ci: switch CodeQL to advanced setup with manual trigger by Akshay Dongare (@Akshay-Dongare) in #164
ci(release): switch to human-in-the-loop PyPI publish via draft releases by Akshay Dongare (@Akshay-Dongare) in #162
Delete .github/workflows/codeql.yml by Akshay Dongare (@Akshay-Dongare) in #165
ci: add PR title conventional commits lint by Akshay Dongare (@Akshay-Dongare) in #163
build(deps): bump the python-deps group with 5 updates by Dependabot (@dependabot)[bot] in #166
ci(deps): pin GitHub Actions to full commit SHAs by Akshay Dongare (@Akshay-Dongare) in #167
test: enable has_structured_output and supports_image_inputs in standard test suites by Akshay Dongare (@Akshay-Dongare) in #168
test: enable supports_json_mode in standard test suites by Akshay Dongare (@Akshay-Dongare) in #169
test: rename factory helpers to avoid PytestReturnNotNoneWarning by Akshay Dongare (@Akshay-Dongare) in #170
ci(pr-title-lint): remove warn-on-bypass job and bypass mechanism by Akshay Dongare (@Akshay-Dongare) in #171
docs: replace generic wrapper descriptions in module docstrings by Akshay Dongare (@Akshay-Dongare) in #172
docs: fix stale and inaccurate docstrings across chat model and embedding modules by Akshay Dongare (@Akshay-Dongare) in #173
refactor: hoist litellm import to module level in _create_retry_decor… by Akshay Dongare (@Akshay-Dongare) in #174
refactor: remove unreachable ImportError guard in validate_environment by Akshay Dongare (@Akshay-Dongare) in #175
ci: pin "CI" github workflow steps to full commit SHAs by Akshay Dongare (@Akshay-Dongare) in #176
docs: port advanced features section to docs.langchain.com by Akshay Dongare (@Akshay-Dongare) in #177
docs: rm -rf docs because ported to https://docs.langchain.com/ by Akshay Dongare (@Akshay-Dongare) in #178
build(deps): bump aiohttp from 3.13.4 to 3.14.0 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #180
build(deps): bump the github-actions group with 2 updates by Dependabot (@dependabot)[bot] in #181
build(deps): bump the python-deps group with 3 updates by Dependabot (@dependabot)[bot] in #182
ci: add PR labeler for CC type and breaking labels by Akshay Dongare (@Akshay-Dongare) in #183
refactor: add missing type annotations and fix mypy override errors by Akshay Dongare (@Akshay-Dongare) in #185
chore(tests): remove unused imports from unit test files by Akshay Dongare (@Akshay-Dongare) in #188
build(deps): bump the python-deps group with 3 updates by Dependabot (@dependabot)[bot] in #190
feat(langchain-litellm): record package version in model metadata by Mason Daugherty (@mdrxy) in #192
build(deps): bump tornado from 6.5.5 to 6.5.6 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #191
chore(main): release langchain-litellm 0.7.0 by @github-actions[bot] in #193

Full Changelog: langchain-litellm==0.6.6...langchain-litellm==0.7.0

Open source →

Additional notes

0.7.0 (2026-06-15)

Features

langchain-litellm: record package version in model metadata (#192) (1c0dcf3)

What's Changed

ci: switch CodeQL to advanced setup with manual trigger by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/164
ci(release): switch to human-in-the-loop PyPI publish via draft releases by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/162
Delete .github/workflows/codeql.yml by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/165
ci: add PR title conventional commits lint by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/163
build(deps): bump the python-deps group with 5 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/166
ci(deps): pin GitHub Actions to full commit SHAs by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/167
test: enable has_structured_output and supports_image_inputs in standard test suites by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/168
test: enable supports_json_mode in standard test suites by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/169
test: rename factory helpers to avoid PytestReturnNotNoneWarning by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/170
ci(pr-title-lint): remove warn-on-bypass job and bypass mechanism by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/171
docs: replace generic wrapper descriptions in module docstrings by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/172
docs: fix stale and inaccurate docstrings across chat model and embedding modules by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/173
refactor: hoist litellm import to module level in _create_retry_decor… by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/174
refactor: remove unreachable ImportError guard in validate_environment by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/175
ci: pin "CI" github workflow steps to full commit SHAs by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/176
docs: port advanced features section to docs.langchain.com by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/177
docs: rm -rf docs because ported to https://docs.langchain.com/ by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/178
build(deps): bump aiohttp from 3.13.4 to 3.14.0 in the python-security group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/180
build(deps): bump the github-actions group with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/181
build(deps): bump the python-deps group with 3 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/182
ci: add PR labeler for CC type and breaking labels by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/183
refactor: add missing type annotations and fix mypy override errors by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/185
chore(tests): remove unused imports from unit test files by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/188
build(deps): bump the python-deps group with 3 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/190
feat(langchain-litellm): record package version in model metadata by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/192
build(deps): bump tornado from 6.5.5 to 6.5.6 in the python-security group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/191
chore(main): release langchain-litellm 0.7.0 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/193

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/langchain-litellm==0.6.6...langchain-litellm==0.7.0

Additional notes

langchain-litellm: v0.7.0

Compare

Choose a tag to compare

Open source →

0.6.621 May 20263 sources agree

bug when api_key is provided but not passed and is overridden by environment variables

0.6.6 (2026-05-21)

Bug Fixes

bug when api_key is provided but not passed and is overridden by environment variables (#134) (1b12aeb)
populate model_provider in response_metadata and ls_provider in _get_ls_params (#152) (#158) (91004e2)
remove global litellm module mutations from _client_params (#132) (#161) (a7ca120)
strip thinking/redacted_thinking blocks from messages sent to non-Anthropic providers (#159) (ce809af)

What's Changed

build(deps): bump langchain-core from 1.3.0 to 1.3.3 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #143
build(deps): bump the github-actions group with 2 updates by Dependabot (@dependabot)[bot] in #144
build(deps): bump the python-deps group with 13 updates by Dependabot (@dependabot)[bot] in #145
Bump test group dependencies by Akshay Dongare (@Akshay-Dongare) in #146
Beautify the pull request title for release-please by Akshay Dongare (@Akshay-Dongare) in #147
build(deps): bump the python-security group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #148
build(deps): bump the python-deps group across 1 directory with 6 updates by Dependabot (@dependabot)[bot] in #150
build(deps): bump langsmith from 0.7.31 to 0.8.0 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #151
fix: bug when api_key is provided but not passed and is overridden by environment variables by Matt Nitzken (@mjnitz02) in #134
chore: add .gitattributes and renormalize line endings by Akshay Dongare (@Akshay-Dongare) in #154
ci(infra): harden Dependabot version-bound preservation by Mason Daugherty (@mdrxy) in #155
build(deps): bump the python-deps group with 7 updates by Dependabot (@dependabot)[bot] in #156
build(deps): bump idna from 3.11 to 3.15 in the python-security group across 1 directory by Dependabot (@dependabot)[bot] in #157
fix: populate model_provider in response_metadata and ls_provider in _get_ls_params (#152) by Akshay Dongare (@Akshay-Dongare) in #158
fix: strip thinking/redacted_thinking blocks from messages sent to non-Anthropic providers by Akshay Dongare (@Akshay-Dongare) in #159
Fix spacing in pull request title pattern by Akshay Dongare (@Akshay-Dongare) in #160
fix: remove global litellm module mutations from _client_params (#132) by Akshay Dongare (@Akshay-Dongare) in #161
chore(main): release langchain-litellm 0.6.6 by @github-actions[bot] in #153

New Contributors

Matt Nitzken (@mjnitz02) made their first contribution in #134

Full Changelog: langchain-litellm==0.6.5...langchain-litellm==0.6.6

Open source →

Additional notes

0.6.6 (2026-05-21)

Bug Fixes

bug when api_key is provided but not passed and is overridden by environment variables (#134) (1b12aeb)
populate model_provider in response_metadata and ls_provider in _get_ls_params (#152) (#158) (91004e2)
remove global litellm module mutations from _client_params (#132) (#161) (a7ca120)
strip thinking/redacted_thinking blocks from messages sent to non-Anthropic providers (#159) (ce809af)

What's Changed

build(deps): bump langchain-core from 1.3.0 to 1.3.3 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/143
build(deps): bump the github-actions group with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/144
build(deps): bump the python-deps group with 13 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/145
Bump test group dependencies by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/146
Beautify the pull request title for release-please by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/147
build(deps): bump the python-security group across 1 directory with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/148
build(deps): bump the python-deps group across 1 directory with 6 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/150
build(deps): bump langsmith from 0.7.31 to 0.8.0 in the python-security group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/151
fix: bug when api_key is provided but not passed and is overridden by environment variables by @mjnitz02 in https://github.com/langchain-ai/langchain-litellm/pull/134
chore: add .gitattributes and renormalize line endings by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/154
ci(infra): harden Dependabot version-bound preservation by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/155
build(deps): bump the python-deps group with 7 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/156
build(deps): bump idna from 3.11 to 3.15 in the python-security group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/157
fix: populate model_provider in response_metadata and ls_provider in _get_ls_params (#152) by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/158
fix: strip thinking/redacted_thinking blocks from messages sent to non-Anthropic providers by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/159
Fix spacing in pull request title pattern by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/160
fix: remove global litellm module mutations from _client_params (#132) by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/161
chore(main): release langchain-litellm 0.6.6 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/153

New Contributors

@mjnitz02 made their first contribution in https://github.com/langchain-ai/langchain-litellm/pull/134

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/langchain-litellm==0.6.5...langchain-litellm==0.6.6

Additional notes

langchain-litellm: v0.6.6

Compare

Choose a tag to compare

Open source →

0.6.58 May 2026

build(deps): bump litellm from 1.81.13 to 1.83.0 in the uv group across 1 directory by Dependabot (@dependabot) [bot] in #129

What's Changed

build(deps): bump litellm from 1.81.13 to 1.83.0 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #129
build(deps): bump the uv group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #130
build(deps): bump langsmith from 0.7.4 to 0.7.31 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #131
build(deps-dev): bump langchain-text-splitters from 0.3.11 to 1.1.2 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #133
build(deps-dev): bump python-dotenv from 1.2.1 to 1.2.2 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #136
fix: patch 3 security alerts (critical+high severity) in litellm by John Kennedy (@jkennedyvz) in #137

New Contributors

John Kennedy (@jkennedyvz) made their first contribution in #137

Full Changelog: langchain-litellm==0.6.4...langchain-litellm==0.6.5

Open source →

Additional notes

What's Changed

build(deps): bump litellm from 1.81.13 to 1.83.0 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/129
build(deps): bump the uv group across 1 directory with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/130
build(deps): bump langsmith from 0.7.4 to 0.7.31 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/131
build(deps-dev): bump langchain-text-splitters from 0.3.11 to 1.1.2 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/133
build(deps-dev): bump python-dotenv from 1.2.1 to 1.2.2 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/136
fix: patch 3 security alerts (critical+high severity) in litellm by @jkennedyvz in https://github.com/langchain-ai/langchain-litellm/pull/137

New Contributors

@jkennedyvz made their first contribution in https://github.com/langchain-ai/langchain-litellm/pull/137

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/langchain-litellm==0.6.4...langchain-litellm==0.6.5

Open source →

Additional notes2 sources agree

Bug Fixes

bump litellm floor to 1.83.14 to clear downstream CVE alerts (a1370a6), closes #141
patch 3 security alerts (critical+high severity) in litellm (#137) (b170dcc)

Open source →

Additional notes

langchain-litellm: v0.6.5

Compare

Choose a tag to compare

Open source →

0.6.43 Apr 20263 sources agree

downgrade tool_choice to auto when thinking is enabled on Claude

0.6.4 (2026-04-03)

Bug Fixes

downgrade tool_choice to auto when thinking is enabled on Claude (#126) (8912d08)
extract reasoning tokens and handle pydantic usage in metadata (#121) (ae18705)
filter tool_use content blocks from AI message dicts (#125) (06c893d)
populate model_name in response_metadata for streaming and router paths (#124) (df1216c)
test (f066e3d)

What's Changed

docs: update platform badge to OS Independent by Akshay Dongare (@Akshay-Dongare) in #97
docs: update Advanced Features --> Embeddings section to include vers… by Akshay Dongare (@Akshay-Dongare) in #98
Bump the uv group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #100
fix: exclude compromised litellm versions from deps by Mason Daugherty (@mdrxy) in #103
chore(main): release 0.6.2 by @github-actions[bot] in #104
Bump requests from 2.32.5 to 2.33.0 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #105
ci: tighten top-level release permissions to contents: read by Mason Daugherty (@mdrxy) in #108
Bump the uv group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #106
fix: set usage_metadata on AIMessage in _create_chat_result by @429er in #102
chore(main): release 0.6.3 by @github-actions[bot] in #109
chore: refine title formatting by Mason Daugherty (@mdrxy) in #110
feat: test by Mason Daugherty (@mdrxy) in #112
Fix/streaming usage metadata all providers by Akshay Dongare (@Akshay-Dongare) in #116
chore: remove unused _is_openai method and _OPENAI_MODELS after strea… by Akshay Dongare (@Akshay-Dongare) in #117
Fix style inconcistencies in test_litellm.py, ensuring all test funct… by Akshay Dongare (@Akshay-Dongare) in #118
chore: format by Mason Daugherty (@mdrxy) in #120
fix: extract reasoning tokens and handle pydantic usage in metadata by Mason Daugherty (@mdrxy) in #121
chore: rename tests by Mason Daugherty (@mdrxy) in #122
chore: fix integration tests by Mason Daugherty (@mdrxy) in #123
fix: populate model_name in response_metadata for streaming and router paths by Mason Daugherty (@mdrxy) in #124
fix: filter tool_use content blocks from AI message dicts by Mason Daugherty (@mdrxy) in #125
build(deps): bump aiohttp from 3.13.3 to 3.13.4 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #114
fix: downgrade tool_choice to auto when thinking is enabled on Claude by Mason Daugherty (@mdrxy) in #126
release(litellm): 0.6.4 by @github-actions[bot] in #113
fix: bump pyproject.toml to 0.6.4 and harden release pipeline by Mason Daugherty (@mdrxy) in #127

Full Changelog: v0.6.1...langchain-litellm==0.6.4

Open source →

Additional notes

0.6.4 (2026-04-03)

Bug Fixes

downgrade tool_choice to auto when thinking is enabled on Claude (#126) (8912d08)
extract reasoning tokens and handle pydantic usage in metadata (#121) (ae18705)
filter tool_use content blocks from AI message dicts (#125) (06c893d)
populate model_name in response_metadata for streaming and router paths (#124) (df1216c)
test (f066e3d)

What's Changed

docs: update platform badge to OS Independent by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/97
docs: update Advanced Features --> Embeddings section to include vers… by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/98
Bump the uv group across 1 directory with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/100
fix: exclude compromised litellm versions from deps by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/103
chore(main): release 0.6.2 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/104
Bump requests from 2.32.5 to 2.33.0 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/105
ci: tighten top-level release permissions to contents: read by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/108
Bump the uv group across 1 directory with 2 updates by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/106
fix: set usage_metadata on AIMessage in _create_chat_result by @429er in https://github.com/langchain-ai/langchain-litellm/pull/102
chore(main): release 0.6.3 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/109
chore: refine title formatting by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/110
feat: test by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/112
Fix/streaming usage metadata all providers by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/116
chore: remove unused _is_openai method and _OPENAI_MODELS after strea… by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/117
Fix style inconcistencies in test_litellm.py, ensuring all test funct… by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/118
chore: format by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/120
fix: extract reasoning tokens and handle pydantic usage in metadata by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/121
chore: rename tests by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/122
chore: fix integration tests by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/123
fix: populate model_name in response_metadata for streaming and router paths by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/124
fix: filter tool_use content blocks from AI message dicts by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/125
build(deps): bump aiohttp from 3.13.3 to 3.13.4 in the uv group across 1 directory by @dependabot[bot] in https://github.com/langchain-ai/langchain-litellm/pull/114
fix: downgrade tool_choice to auto when thinking is enabled on Claude by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/126
release(litellm): 0.6.4 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/113
fix: bump pyproject.toml to 0.6.4 and harden release pipeline by @mdrxy in https://github.com/langchain-ai/langchain-litellm/pull/127

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/v0.6.1...langchain-litellm==0.6.4

Additional notes

langchain-litellm: v0.6.4

Compare

Choose a tag to compare

Open source →

0.6.31 Apr 20263 sources agree

🤖 I have created a release beep boop

0.6.3
(2026-04-01)

Bug Fixes

set usage_metadata on AIMessage in _create_chat_result
(#102)
(75766a0)

This PR was generated with Release
Please. See
documentation.

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Open source →

Additional notes

0.6.3 (2026-04-01)

Bug Fixes

set usage_metadata on AIMessage in _create_chat_result (#102) (75766a0)

What's Changed

Bump requests from 2.32.5 to 2.33.0 in the uv group across 1 directory by Dependabot (@dependabot)[bot] in #105
ci: tighten top-level release permissions to contents: read by Mason Daugherty (@mdrxy) in #108
Bump the uv group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #106
fix: set usage_metadata on AIMessage in _create_chat_result by @429er in #102
chore(main): release 0.6.3 by @github-actions[bot] in #109

Full Changelog: v0.6.2...v0.6.3

Open source →

Additional notes

v0.6.3

Compare

Choose a tag to compare

Open source →

0.6.224 Mar 20263 sources agree

exclude compromised litellm versions from deps

0.6.2 (2026-03-24)

Bug Fixes

exclude compromised litellm versions from deps (#103) (701e152)

What's Changed

docs: update platform badge to OS Independent by Akshay Dongare (@Akshay-Dongare) in #97
docs: update Advanced Features --> Embeddings section to include vers… by Akshay Dongare (@Akshay-Dongare) in #98
Bump the uv group across 1 directory with 2 updates by Dependabot (@dependabot)[bot] in #100
fix: exclude compromised litellm versions from deps by Mason Daugherty (@mdrxy) in #103
chore(main): release 0.6.2 by @github-actions[bot] in #104

Full Changelog: v0.6.1...v0.6.2

Open source →

Additional notes

🤖 I have created a release beep boop

0.6.2
(2026-03-24)

Bug Fixes

exclude compromised litellm versions from deps
(#103)
(701e152)

This PR was generated with Release
Please. See
documentation.

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Open source →

Additional notes

v0.6.2

Compare

Choose a tag to compare

Open source →

0.6.11 Mar 20263 sources agree

clean up build artifacts to unblock PyPI publication

0.6.1 (2026-03-01)

Bug Fixes

clean up build artifacts to unblock PyPI publication (#95) (02e7456)

What's Changed

chore: remove dist folder from version control and add to gitignore by Akshay Dongare (@Akshay-Dongare) in #94
fix: clean up build artifacts to unblock PyPI publication by Akshay Dongare (@Akshay-Dongare) in #95
chore(main): release 0.6.1 by @github-actions[bot] in #96

Full Changelog: v0.6.0...v0.6.1

Open source →

Additional notes

0.6.1 (2026-03-01)

Bug Fixes

clean up build artifacts to unblock PyPI publication (#95) (02e7456)

What's Changed

chore: remove dist folder from version control and add to gitignore by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/94
fix: clean up build artifacts to unblock PyPI publication by @Akshay-Dongare in https://github.com/langchain-ai/langchain-litellm/pull/95
chore(main): release 0.6.1 by @github-actions[bot] in https://github.com/langchain-ai/langchain-litellm/pull/96

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/v0.6.0...v0.6.1

Open source →

Additional notes

v0.6.1: Remove Build Artifacts

Compare

Choose a tag to compare

Open source →

0.5.111 Feb 2026

feat: add configurable timeout and retry logic to LiteLLMOCRLoader by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/68

What's Changed

feat: add configurable timeout and retry logic to LiteLLMOCRLoader by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/68

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.5.0...v0.5.1

Open source →

Additional notes

Features

add configurable timeout and retry logic to LiteLLMOCRLoader (#68) (1c079a5)

Contributors: @Akshay-Dongare

Open source →

Additional notes

v0.5.1: Add configurable timeout and retry logic to LiteLLMOCRLoader

Compare

Choose a tag to compare

Open source →

0.5.09 Feb 2026

LiteLLM OCR Loader by @Bschim in https://github.com/Akshay-Dongare/langchain-litellm/pull/65

What's Changed

LiteLLM OCR Loader by @Bschim in https://github.com/Akshay-Dongare/langchain-litellm/pull/65

New Contributors

@Bschim made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/65

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.4.0...v0.5.0

Open source →

Additional notes

Features

Initial LiteLLM OCR Loader (#65) (ad733f1)

Contributors: @Bschim

Open source →

0.4.02 Feb 2026

Update README with usage examples for Vertex AI by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/48

What's Changed

Update README with usage examples for Vertex AI by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/48
Update model features table in litellm.ipynb by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/49
Fix: Expose cache tokens in streaming responses by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/53
fix: allow reasoning_content and function_call to coexist in message … by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/55
Fix missing usage metadata in streaming responses by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/56
Add support for num_ctx parameter and improve documentation by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/57
Fix request_timeout parameter not being respected by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/58
Fix AttributeError in streaming when chunks lack role or function_cal… by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/59
Add support for gemini-2.5-pro by updating litellm dependency by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/60
Fix multimodal content handling and preserve LiteLLM native format by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/61
Fix structured output crashes: handle tool_choice and Dict arguments by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/62
Add support for base_model parameter to fix fine-tuned Vertex AI models by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/63
Fix streaming crash on Bedrock while preserving OpenAI usage stats by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/64

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.5...v0.4.0

Open source →

Additional notes

Features

Add support for base_model parameter to fix fine-tuned Vertex AI models (#63) (ed37959)
Add support for gemini-2.5-pro by updating litellm dependency (#60) (132d71c)
Add support for num_ctx parameter and improve documentation (#57) (3cbbb43)

Bug Fixes

allow reasoning_content and function_call to coexist in message chunks (#55) (338672d)
Expose cache tokens in streaming responses (#53) (cb02c23)
Fix AttributeError in streaming when chunks lack role or function_call attributes (#59) (62488a0)
Fix missing usage metadata in streaming responses (#56) (ae8b263)
Fix multimodal content handling and preserve LiteLLM native format (#61) (5e2eed1)
Fix request_timeout parameter not being respected (#58) (f231574)
Fix streaming crash on Bedrock while preserving OpenAI usage stats (#64) (28edfda)
Fix structured output crashes: handle tool_choice and Dict arguments (#62) (7793f17)

Contributors: @Akshay-Dongare

0.3.513 Dec 2025

Inject Root Metadata into Delta by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/47

What's Changed

Inject Root Metadata into Delta by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/47

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.4...v0.3.5

Open source →

Additional notes

Bug Fixes

Inject Root Metadata into Delta (#47) (7c239b9)

Contributors: @Akshay-Dongare

Open source →

0.3.413 Dec 2025

Vertex ai grounding metadata by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/45

What's Changed

Vertex ai grounding metadata by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/45

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.3...v0.3.4

Open source →

Additional notes

Features

Add compatibility for Vertex AI specific grounding metadata field in Litellm chat model. (#45) (6368a4d)
Added the handling of 'vertex_ai_grounding_metadata' for provider_specific_fields in LiteLLMRouter. (#45) (0ec4970)

Contributors: @Akshay-Dongare

Open source →

0.3.311 Dec 2025

Fix bug where reason content is not output in invoke by @TBice123123 in https://github.com/Akshay-Dongare/langchain-litellm/pull/25

What's Changed

Fix bug where reason content is not output in invoke by @TBice123123 in https://github.com/Akshay-Dongare/langchain-litellm/pull/25
Add support for provider_specific_field by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/43

New Contributors

@TBice123123 made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/25

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.2...v0.3.3

Open source →

Additional notes

Features

Add "provider_specific_fields" support in chat models and router (#43) (c56a352)

Bug Fixes

Fix bug where reason content is not output in invoke (#25) (dbe0134)

Contributors: @TBice123123, @Akshay-Dongare

Open source →

0.3.220 Nov 2025

Add logprobs to ChatResult for ChatLiteLLM by @allen-cook in https://github.com/Akshay-Dongare/langchain-litellm/pull/40

What's Changed

Add logprobs to ChatResult for ChatLiteLLM by @allen-cook in https://github.com/Akshay-Dongare/langchain-litellm/pull/40
Fix: Preserve litellm_params in ChatLiteLLMRouter by @SamMaggioli in https://github.com/Akshay-Dongare/langchain-litellm/pull/37
add structured output support in ChatLiteLLM by @k4han in https://github.com/Akshay-Dongare/langchain-litellm/pull/36
feat: add support for extra headers in ChatLiteLLM by @k4han in https://github.com/Akshay-Dongare/langchain-litellm/pull/35

New Contributors

@allen-cook made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/40
@SamMaggioli made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/37
@k4han made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/36

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.1...v0.3.2

Open source →

Additional notes

Features

Add logprobs to ChatResult for ChatLiteLLM (#40) (c9e60c7)
add structured output support with schema validation in ChatLiteLLM (#36) (3e6ef60)
add support for extra headers in ChatLiteLLM (#35) (1846a85)

Bug Fixes

filter out None values from params in ChatLiteLLMRouter methods (#37) (6579a2b)

Contributors: @allen-cook, @SamMaggioli, @k4han

Open source →

0.3.120 Nov 2025

Add support for LangChain version 1.0 by @quinlanjager in https://github.com/Akshay-Dongare/langchain-litellm/pull/38

What's Changed

Add support for LangChain version 1.0 by @quinlanjager in https://github.com/Akshay-Dongare/langchain-litellm/pull/38

New Contributors

@quinlanjager made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/38

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.3.0...v0.3.1

Open source →

Additional notes

Features

Add support for LangChain version 1.0 (#38) (03439b1)

Contributors: @quinlanjager

Open source →

0.3.020 Oct 2025

Add Usage Metadata in Streaming Responses by @RheagalFire in https://github.com/Akshay-Dongare/langchain-litellm/pull/34

What's Changed

Add Usage Metadata in Streaming Responses by @RheagalFire in https://github.com/Akshay-Dongare/langchain-litellm/pull/34

New Contributors

@RheagalFire made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/34

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.2.3...v0.3.0

Open source →

Additional notes

Features

Add Usage Metadata in Streaming Responses (#34) (a9abf29)

Contributors: @RheagalFire

Open source →

0.2.325 Sept 2025

fix: ainvoke and astream openai completion api by @maxence-oden in https://github.com/Akshay-Dongare/langchain-litellm/pull/24

What's Changed

fix: ainvoke and astream openai completion api by @maxence-oden in https://github.com/Akshay-Dongare/langchain-litellm/pull/24

New Contributors

@maxence-oden made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/24

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.2.2...v0.2.3

Open source →

Additional notes

Bug Fixes

ainvoke and astream openai completion api (#24) (7675be8)

Contributors: @maxence-oden

Open source →

0.2.211 Jul 2025

Fix tool message parsing by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/11

What's Changed

Fix tool message parsing by @Akshay-Dongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/11
Fix for the Issue:_OPENAI_MODELS is out of date #7 by @AshutoshDongare in https://github.com/Akshay-Dongare/langchain-litellm/pull/12
Fix streaming delta conversion error when chunk is dict by @jeonsworld in https://github.com/Akshay-Dongare/langchain-litellm/pull/14

New Contributors

@Akshay-Dongare made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/11
@AshutoshDongare made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/12
@jeonsworld made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/14

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.2.1...v0.2.2

Open source →

Additional notes

Bug Fixes

Fix for the Issue:_OPENAI_MODELS is out of date #7 (#12) (64cbad1)
Fix tool message conversion (#11) (9fefd6d)
Solves issue #7 (43b6adc)
streaming delta conversion to handle dict types (#14) (8073686)
Update temperature range (846d3d2)

Contributors: @Akshay-Dongare, @AshutoshDongare, @jeonsworld

Open source →

0.2.112 May 2025

fix: fixed streaming with tool calls (#5) by @florianchappaz in https://github.com/Akshay-Dongare/langchain-litellm/pull/6

What's Changed

fix: fixed streaming with tool calls (#5) by @florianchappaz in https://github.com/Akshay-Dongare/langchain-litellm/pull/6

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.2.0...v0.2.1

Open source →

Additional notes

Bug Fixes

fixed streaming with tool calls (#6) (0ddac98)

Contributors: @florianchappaz

Open source →

0.2.026 Apr 2025

feat: added support for ChatLiteLLMRouter (#1) by @florianchappaz in https://github.com/Akshay-Dongare/langchain-litellm/pull/2

What's Changed

feat: added support for ChatLiteLLMRouter (#1) by @florianchappaz in https://github.com/Akshay-Dongare/langchain-litellm/pull/2
ChatLiteLLMRouter docs + litellm update by @florianchappaz in https://github.com/Akshay-Dongare/langchain-litellm/pull/4

New Contributors

@florianchappaz made their first contribution in https://github.com/Akshay-Dongare/langchain-litellm/pull/2

Full Changelog: https://github.com/Akshay-Dongare/langchain-litellm/compare/v0.1.4...v0.2.0

Open source →

Additional notes

Features

added support for ChatLiteLLMRouter (#2) (5cb4bb2)

Contributors: @florianchappaz

Open source →

0.1.45 Apr 2025

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/0.1.3...v0.1.4

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/0.1.3...v0.1.4

Open source →

Additional notes

Release tooling and documentation only, with no functional changes.

Open source →

0.1.35 Apr 2025

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/0.1.2...0.1.3

Full Changelog: https://github.com/langchain-ai/langchain-litellm/compare/0.1.2...0.1.3

Open source →

Additional notes

Features

exporting ChatLiteLLM from __init__.py so you can do from langchain_litellm import ChatLiteLLM (746e14c)

Contributors: @Akshay-Dongare

Open source →

0.1.24 Apr 2025

Full Changelog: https://github.com/langchain-ai/langchain-litellm/commits/0.1.2

Full Changelog: https://github.com/langchain-ai/langchain-litellm/commits/0.1.2

Open source →

Additional notes

Release tooling and documentation only, with no functional changes.

Open source →

0.1.13 Apr 2025

solved poetry dependency issues and rewrote integration tests

Bug Fixes

solved poetry dependency issues and rewrote integration tests (4e92b40)

Contributors: @Akshay-Dongare

Open source →

0.1.03 Apr 2025

### Features * added chat_model code and docs (479fd92) Contributors: @Akshay-Dongare

Features

added chat_model code and docs (479fd92)

Contributors: @Akshay-Dongare

Open source →

Your coding agent can read these notes before it upgrades. Set up the MCP server →

langchain-litellm

Release timeline

Releases

0.11.0 (2026-09-30)

⚠ BREAKING CHANGES

Features

Bug Fixes

What's Changed

Choose a tag to compare

0.10.0 (2026-09-27)

⚠ BREAKING CHANGES

Features

Bug Fixes

What's Changed

New Contributors

Choose a tag to compare

0.9.0 (2026-09-24)

⚠ BREAKING CHANGES

Features

What's Changed

Choose a tag to compare

0.8.1 (2026-09-24)

Bug Fixes

What's Changed

Choose a tag to compare

0.8.0 (2026-09-23)

⚠ BREAKING CHANGES

Features

Bug Fixes

What's Changed

New Contributors

Choose a tag to compare

0.7.2 (2026-09-16)

Bug Fixes

What's Changed

Choose a tag to compare

0.7.1 (2026-08-31)

Bug Fixes

What's Changed

New Contributors

Choose a tag to compare

0.7.0 (2026-06-15)

Features

What's Changed

0.7.0 (2026-06-15)

Features

What's Changed

Choose a tag to compare

0.6.6 (2026-05-21)

Bug Fixes

What's Changed

New Contributors

0.6.6 (2026-05-21)

Bug Fixes

What's Changed

New Contributors

Choose a tag to compare

What's Changed

New Contributors

What's Changed

New Contributors

Bug Fixes

Choose a tag to compare

0.6.4 (2026-04-03)

Bug Fixes

What's Changed

0.6.4 (2026-04-03)

Bug Fixes

What's Changed

Choose a tag to compare

🤖 I have created a release beep boop

Bug Fixes

0.6.3 (2026-04-01)

Bug Fixes

What's Changed

Choose a tag to compare

0.6.2 (2026-03-24)

Bug Fixes

What's Changed

🤖 I have created a release beep boop