NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Go modules · #371 by repository stars
Last release 3 days ago
05 Oct 2026
Ships unpredictably
gaps range from 8 days to 9 months
Some releases are documented
notes for 4 of 16 stable releases
Nothing withdrawn
no release was ever pulled
4 years old
640 releases · first in 2023
One column per quarter.
Nothing published for this version
A patch release that replaces three retired default models, adds typed reasoning content and OpenAI web search, and fixes a number of provider bugs.
A patch release that replaces three retired default models, adds typed
reasoning content and OpenAI web search, and fixes a number of provider bugs.
googleai, vertex: default embedding model is now gemini-embedding-001.
embedding-001 now returns 404. The new model produces 3072-dimension
vectors that are not comparable with the old ones, so indexes built with the
old default must be re-embedded (and fixed-size vector columns resized).
googleai, vertex: default chat model is now gemini-2.5-flash.
gemini-2.0-flash is retired. Gemini 2.5 thinks by default and thinking
counts toward WithMaxTokens; a very small budget can produce no output,
which now returns ErrNoContentInResponse instead of panicking.
Multi-turn tool calling with Gemini 3 is not yet supported (it needs thought
signatures); that is planned for the next release.
anthropic: default model is now claude-sonnet-4-6.
claude-3-5-sonnet-20240620 is retired. Clients that set a model are
unaffected.
Unkeyed struct literals. No exported API was removed, but
llms.ThinkingConfig, llms.ContentChoice, llms.CallOptions and
googleai.Options gained fields, so unkeyed literals of them no longer
compile. Use keyed fields. ContentChoice JSON now includes a Parts key.
Choices[0].StreamingFunc is now returned (with a nil response).temperature/top_p are omitted where the model or thinking mode rejects them.ToolCall JSON round trips keep the function name and arguments;json.UnmarshalTypeError.WithN and WithMinLength no longer set n_ctx and min_p; use WithNCtx and WithMinP.GOOGLE_API_KEY is used only when no other auth option is given.llms.WithWebSearch for OpenAI (#1454).ThinkingContent, RedactedThinkingContent and ContentChoice.Parts.ThinkingConfig.Effort and Display (WithThinkingEffort, WithThinkingDisplay),(*ContentResponse).AssistantMessage for building the next turn of a tool loop.testing/llmtest.TestModel, a conformance check for llms.Model implementations.redacted_thinking blocks; replace the retired default model (#1431).properties (#1446);reasoning_effort only for models that accept it; no temperature for search-preview models.WithHTTPClient; report reasoning support for Gemini 2.5.Full Changelog: v0.1.14...v0.1.15
Nothing published for this version
Nothing published for this version
llms/openai: replace deprecated gpt-4-vision-preview with gpt-4o in TestMultiContentImage by @karsharma10 in #1238
textembedding-gecko with text-embedding-005 in its PaLM Client by @alanrichman in #1292ReasoningContent field to openai GenerateContent functio… by @topjohncian in #1324Full Changelog: v0.1.13...v0.1.14
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Update Milvus SDK deprecation notice
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Removed deprecated VertexAI PaLM support
Full Changelog: v0.1.14-pre.2...v0.1.14-pre.3
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
llms/openai: improve SSE comment handling in streaming responses by @tmc in #1366
ReasoningContent field to openai GenerateContent functio… by @topjohncian in #1324Full Changelog: v0.1.14-pre.1...v0.1.14-pre.2
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
llms/openai: replace deprecated gpt-4-vision-preview with gpt-4o in TestMultiContentImage by @karsharma10 in #1238
textembedding-gecko with text-embedding-005 in its PaLM Client by @alanrichman in #1292Full Changelog: v0.1.13...v0.1.14-pre.1
Nothing published for this version
llms/openai: replace deprecated gpt-4-vision-preview with gpt-4o in TestMultiContentImage by @karsharma10 in #1238
textembedding-gecko with text-embedding-005 in its PaLM Client by @alanrichman in #1292Full Changelog: v0.1.13...v0.1.14-pre.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Reflected max_token field deprecation
We're excited to announce full MongoDB vector store support, one of our most requested features!
Full Changelog: v0.1.12...v0.1.13
Nothing published for this version
Nothing published for this version
openai: reflect max_token field deprecation by @tmc in #1013
Full Changelog: v0.1.12...v0.1.13-pre.1
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →