Packages

@ai-sdk/openai-compatible

3.0.93 · 17 Nov 2026Apache-2.014M / weekRepositoryHomepagenpmAtom feed

Your version

299 stable releases, newest first, by release line. Notes are quoted as the maintainers published them.

  1. 3.x · latest line
  2. 3.0.41

    30 Aug 2026
    • Support text and thinking parts in array-based chat completion content while ignoring unknown part types.

  3. 3.0.37 – 3.0.403 releases that only update dependencies@ai-sdk/provider-utils 5.0.33@ai-sdk/provider 4.0.8Show each
  4. 3.0.36

    24 Aug 2026
    • Send reasoning_effort: "none" when top-level reasoning is disabled

  5. 3.0.35

    22 Aug 2026
    • openai-compatible: Convert video file parts to video_url content parts

  6. 3.0.34

    21 Aug 2026
    • provider/openai-compatible: Report image generation token usage

  7. 3.0.33

    20 Aug 2026
    • openai-compatible: Report truncated chat streams as errors

  8. 3.0.32

    19 Aug 2026
    • Handle empty string tool call IDs

    • Preserve Gemini thought signatures for custom OpenAI-compatible provider names

  9. 3.0.31

    17 Aug 2026
    • 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.

  10. 3.0.29 – 3.0.302 releases that only update dependencies@ai-sdk/provider-utils 5.0.27Show each
  11. 3.0.28

    9 Aug 2026
    • provider/openai-compatible: Clamp outputTokens.text at 0 when a provider reports completion_tokens_details.reasoning_tokens greater than completion_tokens (observed with Baseten serving reasoning models that hit the length stop mid-reasoning). The text share of completion tokens can never be negative; total and reasoning remain as reported by the provider.

  12. 3.0.24 – 3.0.274 releases that only update dependencies@ai-sdk/provider 4.0.7@ai-sdk/provider-utils 5.0.25Show each
  13. 3.0.23

    4 Aug 2026
    • Fix streamed tool calls with non-zero, non-contiguous, reused, or missing indexes.

  14. 3.0.14 – 3.0.229 releases that only update dependencies@ai-sdk/provider-utils 5.0.20@ai-sdk/provider 4.0.5Show each
  15. 3.0.13

    20 Jul 2026
    • openai-compatible: Add an extensible providerOptions type for image model requests and stop forcing response_format

  16. 3.0.12

    17 Jul 2026

    Only updates a dependency:@ai-sdk/provider-utils 5.0.11

  17. 3.0.11

    15 Jul 2026
    • Preserve structured error data from chat completion SSE streams.

  18. 3.0.7 – 3.0.103 releases that only update dependencies@ai-sdk/provider-utils 5.0.10@ai-sdk/provider 4.0.3Show each
  19. 3.0.6

    8 Jul 2026
    • Fix StreamingToolCallTracker finalizing streaming tool calls on parsable partial JSON. Tool calls now only finalize during stream flush, restoring the behavior of #13137: a parsable argument buffer can still be the prefix of a longer argument string, so finalizing early could act on truncated tool inputs.

  20. 3.0.5

    2 Jul 2026
    • Add experimental streaming transcription support for transcription models, including OpenAI gpt-realtime-whisper and xAI WebSocket STT.

  21. 3.0.1 – 3.0.44 releases that only update dependencies@ai-sdk/provider-utils 5.0.4@ai-sdk/provider 4.0.1Show each
  22. 3.0.0

    25 Jun 2026
    • Breaking

      Remove CommonJS exports from all packages. All packages are now ESM-only ("type": "module"). Consumers using require() must switch to ESM import syntax.

    • Breaking

      provider: Add support for provider references and uploading files as supported per provider

    • Breaking

      Start v7 pre-release

  23. 2.x
  24. 2.0.92 – 2.0.943 releases that only update dependencies@ai-sdk/provider-utils 4.0.67Show each
  25. 2.0.91

    6 Nov 2026
    • Avoid empty structured-output answers from Baseten's DeepSeek V4 Pro 0813 when reasoning is enabled. Fall back to JSON prompt instructions while keeping reasoning and SDK output validation, and report the fallback through request transformation warnings.

    • openai,openai-compatible: Surface refusal logprobs and openai-compatible refusals

      @ai-sdk/openai: Chat Completions responses that contain a refusal can return logprobs.refusal next to logprobs.content. The provider previously read only logprobs.content, so providerMetadata.openai.logprobs was empty or missing for refusals. Refusal logprobs are now included in emission order (content logprobs followed by refusal logprobs) in both generateText and streamText.

      Streaming logprobs are now collected across all chunks in streamText: previously only the logprobs of the last streamed chunk were reported. This applies to content logprobs as well, so providerMetadata.openai.logprobs now contains the logprobs of every streamed token instead of only the last chunk's tokens. Explicitly empty logprobs (logprobs: { content:, refusal: null }) are reported as an empty array, as before.

      A refusal that arrives in the same streamed chunk as the finish_reason is now taken into account: such a response is reported with a content-filter finish reason instead of stop.

      @ai-sdk/openai-compatible: Chat Completions responses that carry message.refusal (non-streaming) or delta.refusal (streaming) were ignored, so providers such as OpenAI served through createOpenAICompatible lost the refusal text and reported an empty stop response. Refusals are now returned as text content after any content text, and a non-empty refusal with a stop finish reason is reported as a content-filter finish reason. This also applies when the refusal and the finish reason arrive in the same streamed chunk. Non-string refusals (for example refusal: false from servers that use booleans) are ignored instead of failing the response.

  26. 2.0.90

    3 Nov 2026
    • openai-compatible: Preserve string tool call arguments and safely serialize BigInt and circular structures in tool calls and outputs

  27. 2.0.89

    2 Nov 2026
    • openai-compatible: Send reasoningEffort: null as "reasoning_effort": null, and stop an unset reasoningEffort from overwriting a raw reasoning_effort provider option

  28. 2.0.88

    29 Oct 2026

    Only updates a dependency:@ai-sdk/provider-utils 4.0.63

  29. 2.0.87

    21 Oct 2026
    • Defer validation schema construction until first use to reduce provider import time.

  30. 2.0.86

    19 Oct 2026
    • openai-compatible, groq, alibaba: Put reasoning before text in doGenerate

  31. 2.0.85

    16 Oct 2026
    • openai-compatible: Support URL-based image responses

      Add opt-in URL validation and credential origin handling to getFromApi for URL-based image responses. Validate redirect targets, strip unsafe request headers, and drop credentials on cross-origin redirects while preserving existing unvalidated requests.

A whole project to check? Paste its package.json in the dependency check: every deprecated, outdated or vulnerable dependency, with what to upgrade to. Free, no account.