NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
npm, #3484 most downloaded on npm
Latest stable
4.0.114
Last release 7 days ago
10 Nov 2026
Ships on a steady schedule
a new release about every 8 days
Some releases are documented
notes for 31 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
3 years old
795 releases · first in 2024
Drag across the ruler to compare, or tap a release to open it.
### Patch Changes - Updated dependencies [2ebc78b] - Updated dependencies [cb4fa52] - @ai-sdk/provider-utils@5.0.80 - @ai-sdk/provider@4.0.38
maxOutputTokens and reasoning to experimental decision calls and forward them to language-model-backed decisions. Language-model adapters keep reasoning disabled when omitted. Native decision providers warn when either setting is unsupported.Your coding agent can read these notes before it upgrades. Set up the MCP server →
ede5b89 : chore: migrate package builds from tsup to tsdown
c2511c1 : fix: use standards-compliant User-Agent header
Updated dependencies [ e3605f6 ]
Updated dependencies [ af9597b ]
OpenaiResponsesProviderMetadataOpenaiResponsesReasoningProviderMetadataAzureResponsesProviderMetadataAzureResponsesReasoningProviderMetadataUpdated dependencies [ 9667bf2 ]
Updated dependencies [ fe07867 ]
9184a35: Add ordered text, file, and JSON parts to experimental decision state. Support image input in OpenAI Decisions and language model adapters. G
9184a35: Add ordered text, file, and JSON parts to experimental decision state. Support image input in OpenAI Decisions and language model adapters. Gateway retains its existing string and JSON request format and rejects files, as does TypeSafe AI.
Arrays passed directly as state now contain decision state parts. Wrap JSON arrays in an object or a json part.
Normalize all public state forms into an array of parts before calling decision providers. Providers receive text and JSON objects as text and json parts.
Serialize decision file bytes as base64 in OpenTelemetry state attributes. Model-call spans record normalized state parts, while outer spans retain public state inputs.
Preserve native JSON state in TypeSafe AI when state contains one JSON part. Label shared state in language-model decision prompts.
Updated dependencies [9184a35]
Updated dependencies [ 2973485 ]
Updated dependencies [ 0455398 ]
Updated dependencies [ 215b25e ]
Updated dependencies [ 11b4c2d ]
Updated dependencies [ 0c2a9de ]
### Patch Changes - Updated dependencies [1b11dcb] - @ai-sdk/provider-utils@4.0.6 - @ai-sdk/provider@3.0.3
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
5700d12: Centralize Parallel decision model IDs, supported question types, capabilities, and provider option metadata in a model catalog.
code_interpreter annotations and add typed providerMetadata
code_interpreter annotation types with the official spec.doGenerate and doStream stay in sync.OpenaiResponsesTextProviderMetadataOpenaiResponsesSourceDocumentProviderMetadataAzureResponsesTextProviderMetadataAzureResponsesSourceDocumentProviderMetadataUpdated dependencies [ a4ba394 ]
Updated dependencies [ e5a22f0 ]
image-url tool result content type in OpenAI Responses API conversionUpdated dependencies [ 6bcc0f8 ]
Patch Changes Updated dependencies [9184a35] @ai-sdk/provider@4.0.25 @ai-sdk/provider-utils@5.0.57
@ai-sdk/openai-compatible@3.0.92
790b217: Add a Tencent language and video model catalog for model ID suggestions, backend routing policies, capabilities, and provider option metadata
@latestUpdated dependencies [ 3e125ba ]
c6db718: Add the ScaleDown provider with support for choice, boolean, and score questions through the Decisions API, and text and text-list extraction
0c3b58b: fix(provider): add specificationVersion to ProviderV3
4920119: fix the "incomplete_details" key from nullable to nullish for openai compatibility
0adc679: feat(provider): shared spec v3
92c6241: feat(openai): additional settings for file search tool
88fc415: feat(openai): add the new provider 'apply_patch' tool
817e601: fix(openai); fix url_citation schema in chat api
dae2185: fix(openai): extract meta data from first chunk that contains any
046aa3b: feat(provider): speech model v3 spec
f1277fe: feat(provider/openai): send assistant text and tool call parts as reference ids when store: true
8d9e8ad: chore(provider): remove generics from EmbeddingModelV3
Before
model.textEmbeddingModel("my-model-id");
After
model.embeddingModel("my-model-id");
60f4775: fix: remove code for unsuported o1-mini and o1-preview models
9a51b92: support OPENAI_BASE_URL env
d64ece9: enables image_generation capabilities in the Azure provider through the Responses API.
2625a04: feat(openai); update spec for mcp approval
2e86082: feat(provider/openai): OpenAIChatLanguageModelOptions type
import { openai, type OpenAIChatLanguageModelOptions } from "@ai-sdk/openai";
import { generateText } from "ai";
await generateText({
model: openai.chat("gpt-4o"),
prompt: "Invent a new holiday and describe its traditions.",
providerOptions: {
openai: {
user: "user-123",
} satisfies OpenAIChatLanguageModelOptions,
},
});
0877683: feat(provider/openai): support conversations api
d0f1baf: feat(openai): Add support for 'promptCacheRetention: 24h' for gpt5.1 series
831b6cc: feat(openai): adding provider mcp tool for openai
95f65c2: chore: use import * from zod/v4
edc5548: feat(provider/openai): automatically add reasoning.encrypted_content include when store = false
954c356: feat(openai): allow custom names for provider-defined tools
544d4e8: chore(specification): rename v3 provider defined tool to provider tool
77f2b20: enables code_interpreter and file_search capabilities in the Azure provider through the Responses API
0c4822d: feat: EmbeddingModelV3
73d9883: chore(openai): enable strict json by default
d2039d7: feat(provider/openai): add GPT 5.1 Codex Max to OpenAI Responses model IDs list
88edc28: feat (provider/openai): include more image generation response metadata
e8109d3: feat: tool execution approval
ed329cb: feat: Provider-V3
3bd2689: feat: extended token usage
1cad0ab: feat: add provider version to user-agent header
e85fa2f: feat(openai): add sources in web-search actions
423ba08: Set the annotations from the Responses API to doStream
401f561: fix(provider/openai): fix web search tool input types
4122d2a: feat(provider/openai): add gpt-5-codex model id
0153bfa: fix(openai): fix parameter exclusion logic
8dac895: feat: LanguageModelV3
304222e: Add streaming support for apply_patch partial diffs.
23f132b: fix: error schema for Responses API
1d0de66: refactoring(provider/openai): simplify code
000e87b: fix(provider/openai): add providerExecuted flag to tool start chunks
2c0a758: chore(openai): add JSDoc to responses options
1b982e6: feat(openai): preserve file_id when converting file citations
b82987c: feat(openai): support openai code-interpreter annotations
457318b: chore(provider,ai): switch to SharedV3Warning and unified warnings
b681d7d: feat: expose usage tokens for 'generateImage' function
79b4e46: feat(openai): add 'gpt-5.1' modelID
3997a42: feat(provider/openai): local shell tool
348fd10: fix(openai): treat unknown models as reasoning
9061dc0: feat: image editing
fe49278: feat(provider/openai): only send item references for reasoning when store: true
cb4d238: The built in Code Interpreter tool input code is streamed in tool-input-<start/delta/end> chunks.
357cfd7: feat(provider/openai): add new model IDs gpt-image-1-mini, gpt-5-pro, gpt-5-pro-2025-10-06
38a4035: added support for external_web_access parameter on web_search tool
40d5419: feat(openai): add o3-deep-research and o4-mini-deep-research models
366f50b: chore(provider): add deprecated textEmbeddingModel and textEmbedding aliases
2b0caef: feat(provider/openai): preview image generation results
b60d2e2: fix(openai): allow open_page action type url to be nullish
fd47df5: fix(openai): revised_prompt sometimes returns null, causing errors
4616b86: chore: update zod peer depenedency version
7756857: fix(provider/openai): add truncation parameter support for Responses API
cad6445: feat(openai); adding OpenAI's new shell tool
64aa48f: Azure OpenAI enabled web-search-preview
0b9fdd5: fix(provider/openai): end reasoning parts earlier
61c52dc: feat (provider/openai): add gpt-image-1.5 model support
ef739fa: fix(openai): refactor apply-patch tool
3220329: fix openai responses input: process all provider tool outputs (shell/apply_patch) so parallel tool results aren’t dropped and apply_patch outputs are forwarded.
d270a5d: chore(openai): update tests for apply-patch tool to use snapshots
f18ef7f: feat(openai): add gpt-5.2 models
21e20c0: feat(provider): transcription model v3 spec
522f6b8: feat: ImageModelV3
484aa93: Add 'default' as service tier
88574c1: Change isReasoningModel detection from blocklist to allowlist and add override option
68c6187: feat(provider/openai): support file and image tool results
3794514: feat: flexible tool output content support
cbf52cd: feat: expose raw finish reason
10c1322: fix: moved dependency @ai-sdk/test-server to devDependencies
5648ec0: Add GPT-5.2 support for non-reasoning parameters (temperature, topP, logProbs) when reasoningEffort is none.
78f813e: fix(openai): allow temperature etc setting when reasoning effort is none for gpt-5.1
40dc7fa: fix(openai): change find action type to find_in_page action type
0273b74: fix(openai): add support for sources type 'api'
5bf101a: feat(provider/openai): add support for OpenAI xhigh reasoning effort
1bd7d32: feat: tool-specific strict mode
d86b52f: distinguish between OpenAI and Azure in Responses API providerMetadata
95f65c2: chore: load zod schemas lazily
59561f8: fix(openai); fix url_citation schema in chat api
Updated dependencies
@ai-sdk/provider-utils@4.0.0-beta.59
304222e: Add streaming support for apply_patch partial diffs.
2625a04: feat(openai); update spec for mcp approval
cbf52cd: feat: expose raw finish reason
@ai-sdk/provider-utils@4.0.0-beta.56
@ai-sdk/provider-utils@4.0.0-beta.55
@ai-sdk/provider-utils@4.0.0-beta.54
88574c1: Change isReasoningModel detection from blocklist to allowlist and add override option
isReasoningModel detection from blocklist to allowlist and add override option61c52dc: feat (provider/openai): add gpt-image-1.5 model support
@ai-sdk/provider-utils@4.0.0-beta.52
3220329: fix openai responses input: process all provider tool outputs (shell/apply_patch) so parallel tool results aren’t dropped and apply_patch out
@ai-sdk/provider-utils@4.0.0-beta.51
@ai-sdk/provider-utils@4.0.0-beta.50
@ai-sdk/provider-utils@4.0.0-beta.49
@ai-sdk/provider-utils@4.0.0-beta.48
817e601: fix(openai); fix url_citation schema in chat api
40dc7fa: fix(openai): change find action type to find_in_page action type
f18ef7f: feat(openai): add gpt-5.2 models
d2039d7: feat(provider/openai): add GPT 5.1 Codex Max to OpenAI Responses model IDs list
5bf101a: feat(provider/openai): add support for OpenAI xhigh reasoning effort
@ai-sdk/provider-utils@4.0.0-beta.47
@ai-sdk/provider-utils@4.0.0-beta.46
3bd2689: feat: extended token usage
92c6241: feat(openai): additional settings for file search tool
@ai-sdk/provider-utils@4.0.0-beta.44
0153bfa: fix(openai): fix parameter exclusion logic
78f813e: fix(openai): allow temperature etc setting when reasoning effort is none for gpt-5.1
@ai-sdk/provider-utils@4.0.0-beta.43
ef739fa: fix(openai): refactor apply-patch tool
@ai-sdk/provider-utils@4.0.0-beta.42
cad6445: feat(openai); adding OpenAI's new shell tool
b60d2e2: fix(openai): allow open_page action type url to be nullish
1bd7d32: feat: tool-specific strict mode
2c0a758: chore(openai): add JSDoc to responses options
d270a5d: chore(openai): update tests for apply-patch tool to use snapshots
88edc28: feat (provider/openai): include more image generation response metadata
One column per month.