release: 2.50.0

.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "2.49.0"
+  ".": "2.50.0"
 }

.stats.yml

@@ -1,4 +1,4 @@
 configured_endpoints: 278
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai/openai-0650ebae454b4200491980d7663cb7ceabcff90dcb3ed9f67c3bca3e861a2bf6.yml
-openapi_spec_hash: e9576bced964246b7e685a5ad30afffa
-config_hash: 43e311595bcc4fb6b32eba0684de48d8
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/openai/openai-9152683f2fa71f5555cbbb53123ec013b7c83ee6fed4333e76c22a6f48a05f1c.yml
+openapi_spec_hash: c58a8c1bb1f5bb7245323db9437bd67f
+config_hash: ca12d10f1dbe101a81cd0376f868c674

CHANGELOG.md

@@ -1,5 +1,18 @@
 # Changelog
 
+## 2.50.0 (2026-07-28)
+
+Full Changelog: [v2.49.0...v2.50.0](https://github.com/openai/openai-python/compare/v2.49.0...v2.50.0)
+
+### Features
+
+* **api:** transcription model updates ([fd57393](https://github.com/openai/openai-python/commit/fd57393389eac75af08c2e887cb188590448be20))
+
+
+### Bug Fixes
+
+* **audio:** restore transcription keyword overload ([713a262](https://github.com/openai/openai-python/commit/713a2624966c40f4e5b0c20436b8a79aa8383b08))
+
 ## 2.49.0 (2026-07-27)
 
 Full Changelog: [v2.48.0...v2.49.0](https://github.com/openai/openai-python/compare/v2.48.0...v2.49.0)

api.md

@@ -175,6 +175,7 @@ from openai.types.audio import (
     TranscriptionDiarized,
     TranscriptionDiarizedSegment,
     TranscriptionInclude,
+    TranscriptionLanguage,
     TranscriptionSegment,
     TranscriptionStreamEvent,
     TranscriptionTextDeltaEvent,

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openai"
-version = "2.49.0"
+version = "2.50.0"
 description = "The official Python library for the openai API"
 dynamic = ["readme"]
 license = "Apache-2.0"

src/openai/_version.py

@@ -1,4 +1,4 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
 __title__ = "openai"
-__version__ = "2.49.0"  # x-release-please-version
+__version__ = "2.50.0"  # x-release-please-version

src/openai/resources/webhooks/api.md

@@ -14,6 +14,7 @@ from openai.types.webhooks import (
     FineTuningJobCancelledWebhookEvent,
     FineTuningJobFailedWebhookEvent,
     FineTuningJobSucceededWebhookEvent,
+    LiveCallIncomingWebhookEvent,
     RealtimeCallIncomingWebhookEvent,
     ResponseCancelledWebhookEvent,
     ResponseCompletedWebhookEvent,

src/openai/types/audio/__init__.py

@@ -12,6 +12,7 @@
 from .transcription_segment import TranscriptionSegment as TranscriptionSegment
 from .transcription_verbose import TranscriptionVerbose as TranscriptionVerbose
 from .transcription_diarized import TranscriptionDiarized as TranscriptionDiarized
+from .transcription_language import TranscriptionLanguage as TranscriptionLanguage
 from .translation_create_params import TranslationCreateParams as TranslationCreateParams
 from .transcription_stream_event import TranscriptionStreamEvent as TranscriptionStreamEvent
 from .transcription_create_params import TranscriptionCreateParams as TranscriptionCreateParams

src/openai/types/audio/transcription.py

@@ -5,6 +5,7 @@
 
 from ..._utils import PropertyInfo
 from ..._models import BaseModel
+from .transcription_language import TranscriptionLanguage
 
 __all__ = ["Transcription", "Logprob", "Usage", "UsageTokens", "UsageTokensInputTokenDetails", "UsageDuration"]
 
@@ -70,6 +71,13 @@ class Transcription(BaseModel):
     text: str
     """The transcribed text."""
 
+    languages: Optional[List[TranscriptionLanguage]] = None
+    """The languages detected in the audio.
+
+    Returned by `gpt-transcribe`. An empty array indicates that no language could be
+    reliably detected.
+    """
+
     logprobs: Optional[List[Logprob]] = None
     """The log probabilities of the tokens in the transcription.
 

src/openai/types/audio/transcription_create_params.py

@@ -29,7 +29,7 @@ class TranscriptionCreateParamsBase(TypedDict, total=False):
     model: Required[Union[str, AudioModel]]
     """ID of the model to use.
 
-    The options are `gpt-4o-transcribe`, `gpt-4o-mini-transcribe`,
+    The options are `gpt-transcribe`, `gpt-4o-transcribe`, `gpt-4o-mini-transcribe`,
     `gpt-4o-mini-transcribe-2025-12-15`, `whisper-1` (which is powered by our open
     source Whisper V2 model), and `gpt-4o-transcribe-diarize`.
     """
@@ -54,6 +54,12 @@ class TranscriptionCreateParamsBase(TypedDict, total=False):
     not supported when using `gpt-4o-transcribe-diarize`.
     """
 
+    keywords: SequenceNotStr[str]
+    """Words or phrases to guide transcription of the input audio.
+
+    Supported by `gpt-transcribe`.
+    """
+
     known_speaker_names: SequenceNotStr[str]
     """
     Optional list of speaker names that correspond to the audio samples provided in
@@ -78,6 +84,13 @@ class TranscriptionCreateParamsBase(TypedDict, total=False):
     format will improve accuracy and latency.
     """
 
+    languages: SequenceNotStr[str]
+    """
+    Possible languages of the input audio, in
+    [ISO-639-1](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) format.
+    Supported by `gpt-transcribe`.
+    """
+
     prompt: str
     """An optional text to guide the model's style or continue a previous audio
     segment.

src/openai/types/audio/transcription_language.py

@@ -0,0 +1,12 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from ..._models import BaseModel
+
+__all__ = ["TranscriptionLanguage"]
+
+
+class TranscriptionLanguage(BaseModel):
+    """A language detected in transcribed audio."""
+
+    code: str
+    """The code of a language detected in the audio."""

src/openai/types/audio/transcription_text_done_event.py

@@ -4,6 +4,7 @@
 from typing_extensions import Literal
 
 from ..._models import BaseModel
+from .transcription_language import TranscriptionLanguage
 
 __all__ = ["TranscriptionTextDoneEvent", "Logprob", "Usage", "UsageInputTokenDetails"]
 
@@ -60,6 +61,13 @@ class TranscriptionTextDoneEvent(BaseModel):
     type: Literal["transcript.text.done"]
     """The type of the event. Always `transcript.text.done`."""
 
+    languages: Optional[List[TranscriptionLanguage]] = None
+    """The languages detected in the audio.
+
+    Returned by `gpt-transcribe`. An empty array indicates that no language could be
+    reliably detected.
+    """
+
     logprobs: Optional[List[Logprob]] = None
     """The log probabilities of the individual tokens in the transcription.
 

src/openai/types/audio_model.py

@@ -6,6 +6,7 @@
 
 AudioModel: TypeAlias = Literal[
     "whisper-1",
+    "gpt-transcribe",
     "gpt-4o-transcribe",
     "gpt-4o-mini-transcribe",
     "gpt-4o-mini-transcribe-2025-12-15",

src/openai/types/realtime/audio_transcription.py

@@ -1,6 +1,6 @@
 # File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
 
-from typing import Union, Optional
+from typing import List, Union, Optional
 from typing_extensions import Literal
 
 from ..._models import BaseModel
@@ -16,6 +16,12 @@ class AudioTranscription(BaseModel):
     with `gpt-realtime-whisper` in GA Realtime sessions.
     """
 
+    keywords: Optional[List[str]] = None
+    """Words or phrases to guide transcription of the input audio.
+
+    Supported by `gpt-transcribe` and `gpt-live-transcribe`.
+    """
+
     language: Optional[str] = None
     """The language of the input audio.
 
@@ -24,10 +30,19 @@ class AudioTranscription(BaseModel):
     format will improve accuracy and latency.
     """
 
+    languages: Optional[List[str]] = None
+    """
+    Possible languages of the input audio, in
+    [ISO-639-1](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) format.
+    Supported by `gpt-transcribe` and `gpt-live-transcribe`.
+    """
+
     model: Union[
         str,
         Literal[
             "whisper-1",
+            "gpt-transcribe",
+            "gpt-live-transcribe",
             "gpt-4o-mini-transcribe",
             "gpt-4o-mini-transcribe-2025-12-15",
             "gpt-4o-transcribe",
@@ -38,10 +53,10 @@ class AudioTranscription(BaseModel):
     ] = None
     """The model to use for transcription.
 
-    Current options are `whisper-1`, `gpt-4o-mini-transcribe`,
-    `gpt-4o-mini-transcribe-2025-12-15`, `gpt-4o-transcribe`,
-    `gpt-4o-transcribe-diarize`, and `gpt-realtime-whisper`. Use
-    `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
+    Current options are `whisper-1`, `gpt-transcribe`, `gpt-live-transcribe`,
+    `gpt-4o-mini-transcribe`, `gpt-4o-mini-transcribe-2025-12-15`,
+    `gpt-4o-transcribe`, `gpt-4o-transcribe-diarize`, and `gpt-realtime-whisper`.
+    Use `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
     """
 
     prompt: Optional[str] = None

src/openai/types/realtime/audio_transcription_param.py

@@ -5,6 +5,8 @@
 from typing import Union
 from typing_extensions import Literal, TypedDict
 
+from ..._types import SequenceNotStr
+
 __all__ = ["AudioTranscriptionParam"]
 
 
@@ -16,6 +18,12 @@ class AudioTranscriptionParam(TypedDict, total=False):
     with `gpt-realtime-whisper` in GA Realtime sessions.
     """
 
+    keywords: SequenceNotStr[str]
+    """Words or phrases to guide transcription of the input audio.
+
+    Supported by `gpt-transcribe` and `gpt-live-transcribe`.
+    """
+
     language: str
     """The language of the input audio.
 
@@ -24,10 +32,19 @@ class AudioTranscriptionParam(TypedDict, total=False):
     format will improve accuracy and latency.
     """
 
+    languages: SequenceNotStr[str]
+    """
+    Possible languages of the input audio, in
+    [ISO-639-1](https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes) format.
+    Supported by `gpt-transcribe` and `gpt-live-transcribe`.
+    """
+
     model: Union[
         str,
         Literal[
             "whisper-1",
+            "gpt-transcribe",
+            "gpt-live-transcribe",
             "gpt-4o-mini-transcribe",
             "gpt-4o-mini-transcribe-2025-12-15",
             "gpt-4o-transcribe",
@@ -37,10 +54,10 @@ class AudioTranscriptionParam(TypedDict, total=False):
     ]
     """The model to use for transcription.
 
-    Current options are `whisper-1`, `gpt-4o-mini-transcribe`,
-    `gpt-4o-mini-transcribe-2025-12-15`, `gpt-4o-transcribe`,
-    `gpt-4o-transcribe-diarize`, and `gpt-realtime-whisper`. Use
-    `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
+    Current options are `whisper-1`, `gpt-transcribe`, `gpt-live-transcribe`,
+    `gpt-4o-mini-transcribe`, `gpt-4o-mini-transcribe-2025-12-15`,
+    `gpt-4o-transcribe`, `gpt-4o-transcribe-diarize`, and `gpt-realtime-whisper`.
+    Use `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
     """
 
     prompt: str

src/openai/types/realtime/conversation_item_input_audio_transcription_completed_event.py

@@ -5,6 +5,7 @@
 
 from ..._models import BaseModel
 from .log_prob_properties import LogProbProperties
+from ..audio.transcription_language import TranscriptionLanguage
 
 __all__ = [
     "ConversationItemInputAudioTranscriptionCompletedEvent",
@@ -94,5 +95,12 @@ class ConversationItemInputAudioTranscriptionCompletedEvent(BaseModel):
     model's pricing rather than the realtime model's pricing.
     """
 
+    languages: Optional[List[TranscriptionLanguage]] = None
+    """The languages detected in the audio.
+
+    Returned by `gpt-transcribe`. An empty array indicates that no language could be
+    reliably detected.
+    """
+
     logprobs: Optional[List[LogProbProperties]] = None
     """The log probabilities of the transcription."""

src/openai/types/webhooks/__init__.py

@@ -11,6 +11,7 @@
 from .response_failed_webhook_event import ResponseFailedWebhookEvent as ResponseFailedWebhookEvent
 from .eval_run_canceled_webhook_event import EvalRunCanceledWebhookEvent as EvalRunCanceledWebhookEvent
 from .eval_run_succeeded_webhook_event import EvalRunSucceededWebhookEvent as EvalRunSucceededWebhookEvent
+from .live_call_incoming_webhook_event import LiveCallIncomingWebhookEvent as LiveCallIncomingWebhookEvent
 from .response_cancelled_webhook_event import ResponseCancelledWebhookEvent as ResponseCancelledWebhookEvent
 from .response_completed_webhook_event import ResponseCompletedWebhookEvent as ResponseCompletedWebhookEvent
 from .response_incomplete_webhook_event import ResponseIncompleteWebhookEvent as ResponseIncompleteWebhookEvent

src/openai/types/webhooks/live_call_incoming_webhook_event.py

@@ -0,0 +1,55 @@
+# File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+from typing import List, Optional
+from typing_extensions import Literal
+
+from ..._models import BaseModel
+
+__all__ = ["LiveCallIncomingWebhookEvent", "Data", "DataSipHeader"]
+
+
+class DataSipHeader(BaseModel):
+    """A header from the SIP Invite."""
+
+    name: str
+    """Name of the SIP Header."""
+
+    value: str
+    """Value of the SIP Header."""
+
+
+class Data(BaseModel):
+    """Event data payload."""
+
+    session_id: str
+    """The Transceiver `rtc_...` ID of the pending SIP session.
+
+    The same value appears as `call_id` in `realtime.call.incoming`.
+    """
+
+    sip_headers: List[DataSipHeader]
+    """Headers from the SIP Invite."""
+
+
+class LiveCallIncomingWebhookEvent(BaseModel):
+    """Sent when an incoming API SIP session is available for Live acceptance.
+
+    The
+    same pending session can also emit `realtime.call.incoming`; the first
+    successful Realtime or Live accept endpoint selects the runtime surface.
+    """
+
+    id: str
+    """The unique ID of the event."""
+
+    created_at: int
+    """The Unix timestamp (in seconds) of when the event was created."""
+
+    data: Data
+    """Event data payload."""
+
+    type: Literal["live.call.incoming"]
+    """The type of the event. Always `live.call.incoming`."""
+
+    object: Optional[Literal["event"]] = None
+    """The object of the event. Always `event`."""

src/openai/types/webhooks/realtime_call_incoming_webhook_event.py

@@ -22,14 +22,21 @@ class Data(BaseModel):
     """Event data payload."""
 
     call_id: str
-    """The unique ID of this call."""
+    """The Transceiver `rtc_...` ID of the pending SIP session.
+
+    The same value appears as `session_id` in `live.call.incoming`.
+    """
 
     sip_headers: List[DataSipHeader]
     """Headers from the SIP Invite."""
 
 
 class RealtimeCallIncomingWebhookEvent(BaseModel):
-    """Sent when Realtime API Receives a incoming SIP call."""
+    """
+    Sent when an incoming API SIP session is available for Realtime acceptance.
+    The same pending session can also emit `live.call.incoming`; the first
+    successful Realtime or Live accept endpoint selects the runtime surface.
+    """
 
     id: str
     """The unique ID of the event."""