release: 1.79.0

.github/workflows/ci.yml

@@ -30,6 +30,30 @@ jobs:
       - name: Run lints
         run: ./scripts/lint
 
+  upload:
+    if: github.repository == 'stainless-sdks/openai-python'
+    timeout-minutes: 10
+    name: upload
+    permissions:
+      contents: read
+      id-token: write
+    runs-on: depot-ubuntu-24.04
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Get GitHub OIDC Token
+        id: github-oidc
+        uses: actions/github-script@v6
+        with:
+          script: core.setOutput('github_token', await core.getIDToken());
+
+      - name: Upload tarball
+        env:
+          URL: https://pkg.stainless.com/s
+          AUTH: ${{ steps.github-oidc.outputs.github_token }}
+          SHA: ${{ github.sha }}
+        run: ./scripts/utils/upload-artifact.sh
+
   test:
     timeout-minutes: 10
     name: test

.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "1.78.1"
+  ".": "1.79.0"
 }

.stats.yml

@@ -1,4 +1,4 @@
 configured_endpoints: 101
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai%2Fopenai-794a6ed3c3d3d77887564755168056af8a426b17cf1ec721e3a300503dc22a41.yml
-openapi_spec_hash: 25a81c220713cd5b0bafc221d1dfa79a
-config_hash: 0b768ed1b56c6d82816f0fa40dc4aaf5
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai%2Fopenai-262e171d0a8150ea1192474d16ba3afdf9a054b399f1a49a9c9b697a3073c136.yml
+openapi_spec_hash: 33e00a48df8f94c94f46290c489f132b
+config_hash: d8d5fda350f6db77c784f35429741a2e

CHANGELOG.md

@@ -1,5 +1,27 @@
 # Changelog
 
+## 1.79.0 (2025-05-16)
+
+Full Changelog: [v1.78.1...v1.79.0](https://github.com/openai/openai-python/compare/v1.78.1...v1.79.0)
+
+### Features
+
+* **api:** further updates for evals API ([32c99a6](https://github.com/openai/openai-python/commit/32c99a6f5885d4bf3511a7f06b70000edd274301))
+* **api:** manual updates ([25245e5](https://github.com/openai/openai-python/commit/25245e5e3d0713abfb65b760aee1f12bc61deb41))
+* **api:** responses x eval api ([fd586cb](https://github.com/openai/openai-python/commit/fd586cbdf889c9a5c6b9be177ff02fbfffa3eba5))
+* **api:** Updating Assistants and Evals API schemas ([98ba7d3](https://github.com/openai/openai-python/commit/98ba7d355551213a13803f68d5642eecbb4ffd39))
+
+
+### Bug Fixes
+
+* fix create audio transcription endpoint ([e9a89ab](https://github.com/openai/openai-python/commit/e9a89ab7b6387610e433550207a23973b7edda3a))
+
+
+### Chores
+
+* **ci:** fix installation instructions ([f26c5fc](https://github.com/openai/openai-python/commit/f26c5fc85d98d700b68cb55c8be5d15983a9aeaf))
+* **ci:** upload sdks to package manager ([861f105](https://github.com/openai/openai-python/commit/861f1055768168ab04987a42efcd32a07bc93542))
+
 ## 1.78.1 (2025-05-12)
 
 Full Changelog: [v1.78.0...v1.78.1](https://github.com/openai/openai-python/compare/v1.78.0...v1.78.1)

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openai"
-version = "1.78.1"
+version = "1.79.0"
 description = "The official Python library for the openai API"
 dynamic = ["readme"]
 license = "Apache-2.0"

scripts/utils/upload-artifact.sh

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+set -exuo pipefail
+
+RESPONSE=$(curl -X POST "$URL" \
+  -H "Authorization: Bearer $AUTH" \
+  -H "Content-Type: application/json")
+
+SIGNED_URL=$(echo "$RESPONSE" | jq -r '.url')
+
+if [[ "$SIGNED_URL" == "null" ]]; then
+  echo -e "\033[31mFailed to get signed URL.\033[0m"
+  exit 1
+fi
+
+UPLOAD_RESPONSE=$(tar -cz . | curl -v -X PUT \
+  -H "Content-Type: application/gzip" \
+  --data-binary @- "$SIGNED_URL" 2>&1)
+
+if echo "$UPLOAD_RESPONSE" | grep -q "HTTP/[0-9.]* 200"; then
+  echo -e "\033[32mUploaded build to Stainless storage.\033[0m"
+  echo -e "\033[32mInstallation: pip install 'https://pkg.stainless.com/s/openai-python/$SHA'\033[0m"
+else
+  echo -e "\033[31mFailed to upload artifact.\033[0m"
+  exit 1
+fi

src/openai/_version.py

@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "openai"
-__version__ = "1.78.1"  # x-release-please-version
+__version__ = "1.79.0"  # x-release-please-version

src/openai/resources/audio/transcriptions.py

@@ -57,6 +57,7 @@ def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         response_format: Union[Literal["json"], NotGiven] = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
@@ -77,6 +78,7 @@ def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         response_format: Literal["verbose_json"],
         language: str | NotGiven = NOT_GIVEN,
@@ -97,6 +99,7 @@ def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         response_format: Literal["text", "srt", "vtt"],
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
@@ -118,6 +121,7 @@ def create(
         file: FileTypes,
         model: Union[str, AudioModel],
         stream: Literal[True],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -152,6 +156,11 @@ def create(
 
               Note: Streaming is not supported for the `whisper-1` model and will be ignored.
 
+          chunking_strategy: Controls how the audio is cut into chunks. When set to `"auto"`, the server
+              first normalizes loudness and then uses voice activity detection (VAD) to choose
+              boundaries. `server_vad` object can be provided to tweak VAD detection
+              parameters manually. If unset, the audio is transcribed as a single block.
+
           include: Additional information to include in the transcription response. `logprobs` will
               return the log probabilities of the tokens in the response to understand the
               model's confidence in the transcription. `logprobs` only works with
@@ -200,6 +209,7 @@ def create(
         file: FileTypes,
         model: Union[str, AudioModel],
         stream: bool,
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -234,6 +244,11 @@ def create(
 
               Note: Streaming is not supported for the `whisper-1` model and will be ignored.
 
+          chunking_strategy: Controls how the audio is cut into chunks. When set to `"auto"`, the server
+              first normalizes loudness and then uses voice activity detection (VAD) to choose
+              boundaries. `server_vad` object can be provided to tweak VAD detection
+              parameters manually. If unset, the audio is transcribed as a single block.
+
           include: Additional information to include in the transcription response. `logprobs` will
               return the log probabilities of the tokens in the response to understand the
               model's confidence in the transcription. `logprobs` only works with
@@ -281,6 +296,7 @@ def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -299,6 +315,7 @@ def create(
             {
                 "file": file,
                 "model": model,
+                "chunking_strategy": chunking_strategy,
                 "include": include,
                 "language": language,
                 "prompt": prompt,
@@ -357,26 +374,90 @@ async def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
-        response_format: Union[Literal["json"], NotGiven] = NOT_GIVEN,
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
+        include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
+        response_format: Union[Literal["json"], NotGiven] = NOT_GIVEN,
+        stream: Optional[Literal[False]] | NotGiven = NOT_GIVEN,
         temperature: float | NotGiven = NOT_GIVEN,
-        include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         timestamp_granularities: List[Literal["word", "segment"]] | NotGiven = NOT_GIVEN,
         # Use the following arguments if you need to pass additional parameters to the API that aren't available via kwargs.
         # The extra values given here take precedence over values defined on the client or passed to this method.
         extra_headers: Headers | None = None,
         extra_query: Query | None = None,
         extra_body: Body | None = None,
         timeout: float | httpx.Timeout | None | NotGiven = NOT_GIVEN,
-    ) -> Transcription: ...
+    ) -> TranscriptionCreateResponse:
+        """
+        Transcribes audio into the input language.
+
+        Args:
+          file:
+              The audio file object (not file name) to transcribe, in one of these formats:
+              flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.
+
+          model: ID of the model to use. The options are `gpt-4o-transcribe`,
+              `gpt-4o-mini-transcribe`, and `whisper-1` (which is powered by our open source
+              Whisper V2 model).
+
+          chunking_strategy: Controls how the audio is cut into chunks. When set to `"auto"`, the server
+              first normalizes loudness and then uses voice activity detection (VAD) to choose
+              boundaries. `server_vad` object can be provided to tweak VAD detection
+              parameters manually. If unset, the audio is transcribed as a single block.
+
+          include: Additional information to include in the transcription response. `logprobs` will
+              return the log probabilities of the tokens in the response to understand the
+              model's confidence in the transcription. `logprobs` only works with
+              response_format set to `json` and only with the models `gpt-4o-transcribe` and
+              `gpt-4o-mini-transcribe`.
+
+          language: The language of the input audio. Supplying the input language in
+              [ISO-639-1](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) (e.g. `en`)
+              format will improve accuracy and latency.
+
+          prompt: An optional text to guide the model's style or continue a previous audio
+              segment. The
+              [prompt](https://platform.openai.com/docs/guides/speech-to-text#prompting)
+              should match the audio language.
+
+          response_format: The format of the output, in one of these options: `json`, `text`, `srt`,
+              `verbose_json`, or `vtt`. For `gpt-4o-transcribe` and `gpt-4o-mini-transcribe`,
+              the only supported format is `json`.
+
+          stream: If set to true, the model response data will be streamed to the client as it is
+              generated using
+              [server-sent events](https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format).
+              See the
+              [Streaming section of the Speech-to-Text guide](https://platform.openai.com/docs/guides/speech-to-text?lang=curl#streaming-transcriptions)
+              for more information.
+
+              Note: Streaming is not supported for the `whisper-1` model and will be ignored.
+
+          temperature: The sampling temperature, between 0 and 1. Higher values like 0.8 will make the
+              output more random, while lower values like 0.2 will make it more focused and
+              deterministic. If set to 0, the model will use
+              [log probability](https://en.wikipedia.org/wiki/Log_probability) to
+              automatically increase the temperature until certain thresholds are hit.
+
+          timestamp_granularities: The timestamp granularities to populate for this transcription.
+              `response_format` must be set `verbose_json` to use timestamp granularities.
+              Either or both of these options are supported: `word`, or `segment`. Note: There
+              is no additional latency for segment timestamps, but generating word timestamps
+              incurs additional latency.
+
+          extra_headers: Send extra headers
+
+          extra_query: Add additional query parameters to the request
+          """
 
     @overload
     async def create(
         self,
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         response_format: Literal["verbose_json"],
         language: str | NotGiven = NOT_GIVEN,
@@ -397,6 +478,7 @@ async def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         response_format: Literal["text", "srt", "vtt"],
         language: str | NotGiven = NOT_GIVEN,
@@ -418,6 +500,7 @@ async def create(
         file: FileTypes,
         model: Union[str, AudioModel],
         stream: Literal[True],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -452,6 +535,11 @@ async def create(
 
               Note: Streaming is not supported for the `whisper-1` model and will be ignored.
 
+          chunking_strategy: Controls how the audio is cut into chunks. When set to `"auto"`, the server
+              first normalizes loudness and then uses voice activity detection (VAD) to choose
+              boundaries. `server_vad` object can be provided to tweak VAD detection
+              parameters manually. If unset, the audio is transcribed as a single block.
+
           include: Additional information to include in the transcription response. `logprobs` will
               return the log probabilities of the tokens in the response to understand the
               model's confidence in the transcription. `logprobs` only works with
@@ -500,6 +588,7 @@ async def create(
         file: FileTypes,
         model: Union[str, AudioModel],
         stream: bool,
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -534,6 +623,11 @@ async def create(
 
               Note: Streaming is not supported for the `whisper-1` model and will be ignored.
 
+          chunking_strategy: Controls how the audio is cut into chunks. When set to `"auto"`, the server
+              first normalizes loudness and then uses voice activity detection (VAD) to choose
+              boundaries. `server_vad` object can be provided to tweak VAD detection
+              parameters manually. If unset, the audio is transcribed as a single block.
+
           include: Additional information to include in the transcription response. `logprobs` will
               return the log probabilities of the tokens in the response to understand the
               model's confidence in the transcription. `logprobs` only works with
@@ -581,6 +675,7 @@ async def create(
         *,
         file: FileTypes,
         model: Union[str, AudioModel],
+        chunking_strategy: Optional[transcription_create_params.ChunkingStrategy] | NotGiven = NOT_GIVEN,
         include: List[TranscriptionInclude] | NotGiven = NOT_GIVEN,
         language: str | NotGiven = NOT_GIVEN,
         prompt: str | NotGiven = NOT_GIVEN,
@@ -599,6 +694,7 @@ async def create(
             {
                 "file": file,
                 "model": model,
+                "chunking_strategy": chunking_strategy,
                 "include": include,
                 "language": language,
                 "prompt": prompt,

src/openai/resources/embeddings.py

@@ -66,11 +66,12 @@ def create(
           input: Input text to embed, encoded as a string or array of tokens. To embed multiple
               inputs in a single request, pass an array of strings or array of token arrays.
               The input must not exceed the max input tokens for the model (8192 tokens for
-              `text-embedding-ada-002`), cannot be an empty string, and any array must be 2048
+              all embedding models), cannot be an empty string, and any array must be 2048
               dimensions or less.
               [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken)
-              for counting tokens. Some models may also impose a limit on total number of
-              tokens summed across inputs.
+              for counting tokens. In addition to the per-input token limit, all embedding
+              models enforce a maximum of 300,000 tokens summed across all inputs in a single
+              request.
 
           model: ID of the model to use. You can use the
               [List models](https://platform.openai.com/docs/api-reference/models/list) API to
@@ -181,11 +182,12 @@ async def create(
           input: Input text to embed, encoded as a string or array of tokens. To embed multiple
               inputs in a single request, pass an array of strings or array of token arrays.
               The input must not exceed the max input tokens for the model (8192 tokens for
-              `text-embedding-ada-002`), cannot be an empty string, and any array must be 2048
+              all embedding models), cannot be an empty string, and any array must be 2048
               dimensions or less.
               [Example Python code](https://cookbook.openai.com/examples/how_to_count_tokens_with_tiktoken)
-              for counting tokens. Some models may also impose a limit on total number of
-              tokens summed across inputs.
+              for counting tokens. In addition to the per-input token limit, all embedding
+              models enforce a maximum of 300,000 tokens summed across all inputs in a single
+              request.
 
           model: ID of the model to use. You can use the
               [List models](https://platform.openai.com/docs/api-reference/models/list) API to