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Nothing withdrawn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.
f5fc6ba: Derive known model ID suggestions from each provider's model catalog declarations so catalog updates automatically update autocomplete, while
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
Your coding agent can read these notes before it upgrades. Set up the MCP server →
373b075: Add an AssemblyAI model catalog that drives transcription model ID suggestions, request policies, capability documentation, and provider opti
### Patch Changes - Updated dependencies [2ebc78b] - Updated dependencies [cb4fa52] - @ai-sdk/provider-utils@5.0.80 - @ai-sdk/provider@4.0.38 - @ai-sd
Nothing published for this version
Nothing published for this version
Nothing published for this version
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
b7f3132: Add a Cartesia catalog for speech, transcription, and realtime model IDs, capabilities, and provider options. Document word timestamps only f
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Updated dependencies [ 57086ff ]
Updated dependencies [ e64ed82 ]
### Patch Changes - ai@7.0.68
### Patch Changes - ai@7.0.67
### Patch Changes - Updated dependencies [0782259] - Updated dependencies [2fd1214] - ai@7.0.66
### Patch Changes - Updated dependencies [dc8caae] - Updated dependencies [72ec74f] - Updated dependencies [c5b0515] - ai@7.0.65
f5fc6ba: Derive known model ID suggestions from each provider's model catalog declarations so catalog updates automatically update autocomplete, while
@ai-sdk/openai-compatible@3.0.92
@ai-sdk/openai-compatible@3.0.91
Nothing published for this version
### Patch Changes - Updated dependencies [31205a4] - @ai-sdk/provider-utils@4.0.46
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
203a0e0: Centralize Fish Audio speech and transcription model IDs, timestamp defaults, loudness normalization support, and provider option metadata in
Patch Changes Updated dependencies [9184a35] @ai-sdk/provider-utils@5.0.57
f5fc6ba: Derive known model ID suggestions from each provider's model catalog declarations so catalog updates automatically update autocomplete, while
@ai-sdk/openai-compatible@3.0.92
@ai-sdk/openai-compatible@3.0.91
ede5b89 : chore: migrate package builds from tsup to tsdown
Nothing published for this version
Nothing published for this version
Updated dependencies [ 6400cf8 ]
9345eac : Refactor Z.AI model ID logic into a model catalog that declares capabilities and provider options, and generate provider capability document
Updated dependencies [ 36802d4 ]
Updated dependencies [ fa15312 ]
Updated dependencies [ e5a22f0 ]
7716e18 : Support promises and synchronous or asynchronous callbacks in provider API key and token settings. Resolve credentials per authenticated req
7716e18: Support promises and synchronous or asynchronous callbacks in provider API key and token settings. Resolve credentials per authenticated request while preserving existing authentication behavior and synchronous credential serialization.
The experimental OpenAI Live getServerWebSocketConfig() method can now resolve asynchronously. Await its result before using the connection settings.
Updated dependencies [7716e18]
Updated dependencies [5c247e9]
Updated dependencies [dcd2f8d]
Updated dependencies [ 61a315e ]
Updated dependencies [ 57086ff ]
ebf0bd5 : fix(togetherai): forward JSON schemas for DeepSeek V4 Flash structured outputs
### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
05bf552: Add a TypeSafe AI decision model catalog for model ID suggestions, supported question types, image inputs, and call-level and file-part provi
### Patch Changes - Updated dependencies [31205a4] - @ai-sdk/provider-utils@4.0.46
Updated dependencies [ 9e1d1b2 ]
Updated dependencies [ b74971f ]
53bdfa5: fix(openai): allow null/undefined type in streaming tool call deltas
Azure AI Foundry and Mistral deployed on Azure omit the type field in
streaming tool_calls deltas. The chat stream parser now accepts a missing
type field (treating it as "function") instead of throwing
InvalidResponseDataError: Expected 'function' type.
Fixes #12770
9435df8: feat(provider/moonshotai): normalize tool schemas for Moonshot's MFJS validator. Tuple items arrays become prefixItems, type next to anyOf mo
items arrays become prefixItems, type next to anyOf moves into the branches, and non-object root schemas fail with a clear client-side error instead of Moonshot's opaque 400. Everything else passes through unchanged; the original schema is still used for AI SDK result validation.5e18272: fix(openai): include reasoning parts without itemId when encrypted_content is present
When providerOptions.openai.itemId is absent on a reasoning content part,
the converter now uses encrypted_content as a fallback instead of silently
skipping the part with a warning. The OpenAI Responses API accepts reasoning
items without an id when encrypted_content is supplied, enabling
multi-turn reasoning even when item IDs are stripped from provider options.
Also makes the id field optional on the OpenAIResponsesReasoning type to
reflect that the API does not require it.
Fixes #12853
ba40fab: fix(fireworks): enable structured outputs so response_format json_schema is forwarded instead of degrading to json_object
phase parameter on Responses API message items. The phase field ('commentary' or 'final_answer') is returned by models like gpt-5.3-codex on assistant message output items and must be preserved when sending follow-up requests. The phase value is available in providerMetadata.openai.phase on text parts and is automatically included on assistant messages sent back to the API.@ai-sdk/openai-compatible@3.0.31
@ai-sdk/openai-compatible@3.0.31
gpt-5.3-codex86892f3: fix(openai-compatible): preserve unmapped usage fields in usage.raw
86892f3: fix(openai-compatible): preserve unmapped usage fields in usage.raw
usage.raw is specified as usage "in the shape that the provider returns",
and the chat model already parsed the usage object loosely so that extra
top-level fields survived. The nested prompt_tokens_details and
completion_tokens_details objects were still strict, so anything a provider
reported inside them was dropped — which is where providers put their most
distinguishing detail. The existing "should preserve extra usage fields"
fixture was itself losing audio_tokens, image_tokens and text_tokens
this way.
Both nested objects are now parsed loosely, as is the completion model's usage
schema, which was strict throughout despite feeding the same raw field.
Only usage.raw changes. The mapped token counts are unaffected.
Updated dependencies [ 9667bf2 ]
Updated dependencies [ b74971f ]
@ai-sdk/openai-compatible@3.0.66
3494c16: Add a Topaz image and video model catalog for model ID suggestions, API aliases, capabilities, polling defaults, and provider option metadata
### Patch Changes - Updated dependencies [7168375] - @ai-sdk/provider@3.0.8 - @ai-sdk/provider-utils@4.0.14
f5fc6ba: Derive known model ID suggestions from each provider's model catalog declarations so catalog updates automatically update autocomplete, while
heygen.decisionModel('heygen-decision-1') and heygen-decision-1-turbo, including image and video inputs### Patch Changes - Updated dependencies [6d50947] - @ai-sdk/provider-utils@5.0.81
One column per month.