NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
npm · #3484 most downloaded on npm
Last release 4 days ago
01 Oct 2026
Ships on a steady schedule
a new release about every 8 days
Some releases are documented
notes for 32 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
740 releases · first in 2024
### Patch Changes - Updated dependencies [0bd5bc6] - @ai-sdk/provider@1.0.12 - @ai-sdk/provider-utils@2.1.14
2e1101a: feat (provider/openai): pdf input support
One column per month.
523f128: feat (provider/openai): add strictSchemas option to responses model
### Patch Changes - Updated dependencies [1531959] - @ai-sdk/provider-utils@2.1.12
e3a389e: feat (provider/openai): support responses api
e1d3d42: feat (ai): expose raw response body in generateText and generateObject
ede6d1b: feat (provider/azure): Add Azure image model support
d8216f8: feat (provider/openai): add gpt-4.5-preview to model id set
### Patch Changes - Updated dependencies [ddf9740] - @ai-sdk/provider@1.0.9 - @ai-sdk/provider-utils@2.1.10
### Patch Changes - Updated dependencies [2761f06] - @ai-sdk/provider@1.0.8 - @ai-sdk/provider-utils@2.1.9
ea159cb: chore (provider/openai): remove default streaming simulation for o1
### Patch Changes - Updated dependencies [2e898b4] - @ai-sdk/provider-utils@2.1.8
### Patch Changes - Updated dependencies [3ff4ef8] - @ai-sdk/provider-utils@2.1.7
c55b81a: fix (provider/openai): fix o3-mini streaming
161be90: fix (provider/openai): fix model id typo
0a2f026: feat (provider/openai): add o3-mini
@ai-sdk/openai-compatible@2.0.69
@ai-sdk/openai-compatible@2.0.68
### Patch Changes - Updated dependencies [066206e] - @ai-sdk/provider-utils@2.1.4
### Patch Changes - Updated dependencies [39e5c1f] - @ai-sdk/provider-utils@2.1.3
3a58a2e: feat (ai/core): throw NoImageGeneratedError from generateImage when no predictions are returned.
e7a9ec9: feat (provider-utils): include raw value in json parse results
### Minor Changes - 62ba5ad: release: AI SDK 4.1 ### Patch Changes - Updated dependencies [62ba5ad] - @ai-sdk/provider-utils@2.1.0
### Patch Changes - Updated dependencies [00114c5] - @ai-sdk/provider-utils@2.0.8
### Patch Changes - Updated dependencies [31205a4] - @ai-sdk/provider-utils@4.0.46
fe816e4: fix (provider/openai): streamObject with o1
ba62cf2: feat (provider/openai): automatically map maxTokens to max_completion_tokens for reasoning models
### Patch Changes - Updated dependencies - @ai-sdk/provider-utils@2.0.7
f8c6acb: feat (provider/openai): automatically simulate streaming for reasoning models
19a2ce7: feat (ai/core): add aspectRatio and seed options to generateImage
### Patch Changes - Updated dependencies [d300737] - @ai-sdk/harness@1.0.75
### Patch Changes - Updated dependencies [d300737] - @ai-sdk/harness@1.0.75 - @ai-sdk/harness-acp@1.0.13
### Patch Changes - @ai-sdk/harness@1.0.74 - @ai-sdk/harness-acp@1.0.12
HarnessAgent via output property62a9c2a: feat(harness): add support for structured output to HarnessAgent via output property
HarnessAgent via output propertyfinish-step emission timing52bc889: feat(harness): add getPortEndpoint() as more comprehensive replacement to getPortUrl() (now deprecated) in HarnessV1NetworkSandboxSession
getPortEndpoint() as more comprehensive replacement to getPortUrl() (now deprecated) in HarnessV1NetworkSandboxSessionprepareSandboxForHarness() helper69bb613: feat(harness): support request transformations in network sandbox abstraction and use it to apply credential brokering when available
package.json and pnpm-lock.yaml files at build time instead of reading them at runtime to fix runtime errors in certain environments### Patch Changes - Updated dependencies [8d717b3] - @ai-sdk/harness@1.0.71 - @ai-sdk/harness-acp@1.0.8
Updated dependencies [ 14810cb ]
Updated dependencies [ bf29419 ]
Updated dependencies [ 594787e ]
### Patch Changes - Updated dependencies [d300737] - @ai-sdk/harness@1.0.75
### Patch Changes - @ai-sdk/harness@1.0.74
62a9c2a: feat(harness): add support for structured output to HarnessAgent via output property
HarnessAgent via output property80b40aa: feat(harness-cline): implement Cline harness adapter
@ai-sdk/provider-utils@2.0.0-canary.3
@ai-sdk/provider-utils@2.0.0-canary.2
79644e9: chore (provider/openai): remove OpenAI facade
66060f7: chore (release): bump major version to 4.0
0bc4115: feat (provider/openai): support predicted outputs
54a3a59: fix (provider/openai): support object-json mode without schema
3b1b69a: feat: provider-defined tools
741ca51: feat (provider/openai): support mp3 and wav audio inputs
39fccee: feat (provider/openai): provider name can be changed for 3rd party openai compatible providers
3f29c10: feat (provider/openai): support metadata field for distillation
e8aed44: Add OpenAI cached prompt tokens to experimental_providerMetadata for generateText and streamText
5aa576d: feat (provider/openai): support store parameter for distillation
### Patch Changes - Updated dependencies [d595d0d] - @ai-sdk/provider@0.0.24 - @ai-sdk/provider-utils@1.0.20
7efa867: feat (provider/openai): simulated streaming for reasoning models
8132a60: feat (provider/openai): support reasoning token usage and max_completion_tokens
### Patch Changes - Updated dependencies [273f696] - @ai-sdk/provider-utils@1.0.19
Your coding agent can read these notes before it upgrades. Set up the MCP server →