openai 0.30.0 → 0.32.0

data/lib/openai/models/comparison_filter.rb

@@ -10,7 +10,8 @@ module OpenAI
       required :key, String
 
       # @!attribute type
-      #   Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`.
+      #   Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`, `in`,
+      #   `nin`.
       #
       #   - `eq`: equals
       #   - `ne`: not equal
@@ -18,6 +19,8 @@ module OpenAI
       #   - `gte`: greater than or equal
       #   - `lt`: less than
       #   - `lte`: less than or equal
+      #   - `in`: in
+      #   - `nin`: not in
       #
       #   @return [Symbol, OpenAI::Models::ComparisonFilter::Type]
       required :type, enum: -> { OpenAI::ComparisonFilter::Type }
@@ -26,7 +29,7 @@ module OpenAI
       #   The value to compare against the attribute key; supports string, number, or
       #   boolean types.
       #
-      #   @return [String, Float, Boolean]
+      #   @return [String, Float, Boolean, Array<String, Float>]
       required :value, union: -> { OpenAI::ComparisonFilter::Value }
 
       # @!method initialize(key:, type:, value:)
@@ -38,11 +41,12 @@ module OpenAI
       #
       #   @param key [String] The key to compare against the value.
       #
-      #   @param type [Symbol, OpenAI::Models::ComparisonFilter::Type] Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`.
+      #   @param type [Symbol, OpenAI::Models::ComparisonFilter::Type] Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`, `in`, `
       #
-      #   @param value [String, Float, Boolean] The value to compare against the attribute key; supports string, number, or bool
+      #   @param value [String, Float, Boolean, Array<String, Float>] The value to compare against the attribute key; supports string, number, or bool
 
-      # Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`.
+      # Specifies the comparison operator: `eq`, `ne`, `gt`, `gte`, `lt`, `lte`, `in`,
+      # `nin`.
       #
       # - `eq`: equals
       # - `ne`: not equal
@@ -50,6 +54,8 @@ module OpenAI
       # - `gte`: greater than or equal
       # - `lt`: less than
       # - `lte`: less than or equal
+      # - `in`: in
+      # - `nin`: not in
       #
       # @see OpenAI::Models::ComparisonFilter#type
       module Type
@@ -79,8 +85,25 @@ module OpenAI
 
         variant OpenAI::Internal::Type::Boolean
 
+        variant -> { OpenAI::Models::ComparisonFilter::Value::UnionMember3Array }
+
+        module UnionMember3
+          extend OpenAI::Internal::Type::Union
+
+          variant String
+
+          variant Float
+
+          # @!method self.variants
+          #   @return [Array(String, Float)]
+        end
+
         # @!method self.variants
-        #   @return [Array(String, Float, Boolean)]
+        #   @return [Array(String, Float, Boolean, Array<String, Float>)]
+
+        # @type [OpenAI::Internal::Type::Converter]
+        UnionMember3Array =
+          OpenAI::Internal::Type::ArrayOf[union: -> { OpenAI::ComparisonFilter::Value::UnionMember3 }]
       end
     end
   end

data/lib/openai/models/evals/create_eval_completions_run_data_source.rb

@@ -466,6 +466,9 @@ module OpenAI
           #   effort can result in faster responses and fewer tokens used on reasoning in a
           #   response.
           #
+          #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+          #   effort.
+          #
           #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
           optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/evals/run_cancel_response.rb

@@ -320,6 +320,9 @@ module OpenAI
                 #   effort can result in faster responses and fewer tokens used on reasoning in a
                 #   response.
                 #
+                #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+                #   effort.
+                #
                 #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
                 optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 
@@ -661,6 +664,9 @@ module OpenAI
               #   effort can result in faster responses and fewer tokens used on reasoning in a
               #   response.
               #
+              #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+              #   effort.
+              #
               #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
               optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/evals/run_create_params.rb

@@ -232,6 +232,9 @@ module OpenAI
                 #   effort can result in faster responses and fewer tokens used on reasoning in a
                 #   response.
                 #
+                #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+                #   effort.
+                #
                 #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
                 optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 
@@ -589,6 +592,9 @@ module OpenAI
               #   effort can result in faster responses and fewer tokens used on reasoning in a
               #   response.
               #
+              #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+              #   effort.
+              #
               #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
               optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/evals/run_create_response.rb

@@ -320,6 +320,9 @@ module OpenAI
                 #   effort can result in faster responses and fewer tokens used on reasoning in a
                 #   response.
                 #
+                #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+                #   effort.
+                #
                 #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
                 optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 
@@ -661,6 +664,9 @@ module OpenAI
               #   effort can result in faster responses and fewer tokens used on reasoning in a
               #   response.
               #
+              #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+              #   effort.
+              #
               #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
               optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/evals/run_list_response.rb

@@ -320,6 +320,9 @@ module OpenAI
                 #   effort can result in faster responses and fewer tokens used on reasoning in a
                 #   response.
                 #
+                #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+                #   effort.
+                #
                 #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
                 optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 
@@ -661,6 +664,9 @@ module OpenAI
               #   effort can result in faster responses and fewer tokens used on reasoning in a
               #   response.
               #
+              #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+              #   effort.
+              #
               #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
               optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/evals/run_retrieve_response.rb

@@ -320,6 +320,9 @@ module OpenAI
                 #   effort can result in faster responses and fewer tokens used on reasoning in a
                 #   response.
                 #
+                #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+                #   effort.
+                #
                 #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
                 optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 
@@ -665,6 +668,9 @@ module OpenAI
               #   effort can result in faster responses and fewer tokens used on reasoning in a
               #   response.
               #
+              #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+              #   effort.
+              #
               #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
               optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/graders/score_model_grader.rb

@@ -226,6 +226,9 @@ module OpenAI
           #   effort can result in faster responses and fewer tokens used on reasoning in a
           #   response.
           #
+          #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+          #   effort.
+          #
           #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
           optional :reasoning_effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/realtime/audio_transcription.rb

@@ -14,7 +14,8 @@ module OpenAI
 
         # @!attribute model
         #   The model to use for transcription. Current options are `whisper-1`,
-        #   `gpt-4o-transcribe-latest`, `gpt-4o-mini-transcribe`, and `gpt-4o-transcribe`.
+        #   `gpt-4o-mini-transcribe`, `gpt-4o-transcribe`, and `gpt-4o-transcribe-diarize`.
+        #   Use `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
         #
         #   @return [Symbol, OpenAI::Models::Realtime::AudioTranscription::Model, nil]
         optional :model, enum: -> { OpenAI::Realtime::AudioTranscription::Model }
@@ -23,8 +24,8 @@ module OpenAI
         #   An optional text to guide the model's style or continue a previous audio
         #   segment. For `whisper-1`, the
         #   [prompt is a list of keywords](https://platform.openai.com/docs/guides/speech-to-text#prompting).
-        #   For `gpt-4o-transcribe` models, the prompt is a free text string, for example
-        #   "expect words related to technology".
+        #   For `gpt-4o-transcribe` models (excluding `gpt-4o-transcribe-diarize`), the
+        #   prompt is a free text string, for example "expect words related to technology".
         #
         #   @return [String, nil]
         optional :prompt, String
@@ -35,21 +36,22 @@ module OpenAI
         #
         #   @param language [String] The language of the input audio. Supplying the input language in
         #
-        #   @param model [Symbol, OpenAI::Models::Realtime::AudioTranscription::Model] The model to use for transcription. Current options are `whisper-1`, `gpt-4o-tra
+        #   @param model [Symbol, OpenAI::Models::Realtime::AudioTranscription::Model] The model to use for transcription. Current options are `whisper-1`, `gpt-4o-min
         #
         #   @param prompt [String] An optional text to guide the model's style or continue a previous audio
 
         # The model to use for transcription. Current options are `whisper-1`,
-        # `gpt-4o-transcribe-latest`, `gpt-4o-mini-transcribe`, and `gpt-4o-transcribe`.
+        # `gpt-4o-mini-transcribe`, `gpt-4o-transcribe`, and `gpt-4o-transcribe-diarize`.
+        # Use `gpt-4o-transcribe-diarize` when you need diarization with speaker labels.
         #
         # @see OpenAI::Models::Realtime::AudioTranscription#model
         module Model
           extend OpenAI::Internal::Type::Enum
 
           WHISPER_1 = :"whisper-1"
-          GPT_4O_TRANSCRIBE_LATEST = :"gpt-4o-transcribe-latest"
           GPT_4O_MINI_TRANSCRIBE = :"gpt-4o-mini-transcribe"
           GPT_4O_TRANSCRIBE = :"gpt-4o-transcribe"
+          GPT_4O_TRANSCRIBE_DIARIZE = :"gpt-4o-transcribe-diarize"
 
           # @!method self.values
           #   @return [Array<Symbol>]

data/lib/openai/models/reasoning.rb

@@ -10,6 +10,9 @@ module OpenAI
       #   effort can result in faster responses and fewer tokens used on reasoning in a
       #   response.
       #
+      #   Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+      #   effort.
+      #
       #   @return [Symbol, OpenAI::Models::ReasoningEffort, nil]
       optional :effort, enum: -> { OpenAI::ReasoningEffort }, nil?: true
 

data/lib/openai/models/reasoning_effort.rb

@@ -7,6 +7,9 @@ module OpenAI
     # supported values are `minimal`, `low`, `medium`, and `high`. Reducing reasoning
     # effort can result in faster responses and fewer tokens used on reasoning in a
     # response.
+    #
+    # Note: The `gpt-5-pro` model defaults to (and only supports) `high` reasoning
+    # effort.
     module ReasoningEffort
       extend OpenAI::Internal::Type::Enum
 

data/lib/openai/models/vector_store_create_params.rb

@@ -14,6 +14,13 @@ module OpenAI
       #   @return [OpenAI::Models::AutoFileChunkingStrategyParam, OpenAI::Models::StaticFileChunkingStrategyObjectParam, nil]
       optional :chunking_strategy, union: -> { OpenAI::FileChunkingStrategyParam }
 
+      # @!attribute description
+      #   A description for the vector store. Can be used to describe the vector store's
+      #   purpose.
+      #
+      #   @return [String, nil]
+      optional :description, String
+
       # @!attribute expires_after
       #   The expiration policy for a vector store.
       #
@@ -45,12 +52,14 @@ module OpenAI
       #   @return [String, nil]
       optional :name, String
 
-      # @!method initialize(chunking_strategy: nil, expires_after: nil, file_ids: nil, metadata: nil, name: nil, request_options: {})
+      # @!method initialize(chunking_strategy: nil, description: nil, expires_after: nil, file_ids: nil, metadata: nil, name: nil, request_options: {})
       #   Some parameter documentations has been truncated, see
       #   {OpenAI::Models::VectorStoreCreateParams} for more details.
       #
       #   @param chunking_strategy [OpenAI::Models::AutoFileChunkingStrategyParam, OpenAI::Models::StaticFileChunkingStrategyObjectParam] The chunking strategy used to chunk the file(s). If not set, will use the `auto`
       #
+      #   @param description [String] A description for the vector store. Can be used to describe the vector store's p
+      #
       #   @param expires_after [OpenAI::Models::VectorStoreCreateParams::ExpiresAfter] The expiration policy for a vector store.
       #
       #   @param file_ids [Array<String>] A list of [File](https://platform.openai.com/docs/api-reference/files) IDs that

data/lib/openai/models/vector_stores/vector_store_file.rb

@@ -101,7 +101,7 @@ module OpenAI
         # @see OpenAI::Models::VectorStores::VectorStoreFile#last_error
         class LastError < OpenAI::Internal::Type::BaseModel
           # @!attribute code
-          #   One of `server_error` or `rate_limit_exceeded`.
+          #   One of `server_error`, `unsupported_file`, or `invalid_file`.
           #
           #   @return [Symbol, OpenAI::Models::VectorStores::VectorStoreFile::LastError::Code]
           required :code, enum: -> { OpenAI::VectorStores::VectorStoreFile::LastError::Code }
@@ -116,11 +116,11 @@ module OpenAI
           #   The last error associated with this vector store file. Will be `null` if there
           #   are no errors.
           #
-          #   @param code [Symbol, OpenAI::Models::VectorStores::VectorStoreFile::LastError::Code] One of `server_error` or `rate_limit_exceeded`.
+          #   @param code [Symbol, OpenAI::Models::VectorStores::VectorStoreFile::LastError::Code] One of `server_error`, `unsupported_file`, or `invalid_file`.
           #
           #   @param message [String] A human-readable description of the error.
 
-          # One of `server_error` or `rate_limit_exceeded`.
+          # One of `server_error`, `unsupported_file`, or `invalid_file`.
           #
           # @see OpenAI::Models::VectorStores::VectorStoreFile::LastError#code
           module Code

data/lib/openai/resources/audio/transcriptions.rb

@@ -12,7 +12,7 @@ module OpenAI
         #
         # Transcribes audio into the input language.
         #
-        # @overload create(file:, model:, chunking_strategy: nil, include: nil, language: nil, prompt: nil, response_format: nil, temperature: nil, timestamp_granularities: nil, request_options: {})
+        # @overload create(file:, model:, chunking_strategy: nil, include: nil, known_speaker_names: nil, known_speaker_references: nil, language: nil, prompt: nil, response_format: nil, temperature: nil, timestamp_granularities: nil, request_options: {})
         #
         # @param file [Pathname, StringIO, IO, String, OpenAI::FilePart] The audio file object (not file name) to transcribe, in one of these formats: fl
         #
@@ -22,6 +22,10 @@ module OpenAI
         #
         # @param include [Array<Symbol, OpenAI::Models::Audio::TranscriptionInclude>] Additional information to include in the transcription response.
         #
+        # @param known_speaker_names [Array<String>] Optional list of speaker names that correspond to the audio samples provided in
+        #
+        # @param known_speaker_references [Array<String>] Optional list of audio samples (as [data URLs](https://developer.mozilla.org/en-
+        #
         # @param language [String] The language of the input audio. Supplying the input language in [ISO-639-1](htt
         #
         # @param prompt [String] An optional text to guide the model's style or continue a previous audio segment
@@ -34,7 +38,7 @@ module OpenAI
         #
         # @param request_options [OpenAI::RequestOptions, Hash{Symbol=>Object}, nil]
         #
-        # @return [OpenAI::Models::Audio::Transcription, OpenAI::Models::Audio::TranscriptionVerbose]
+        # @return [OpenAI::Models::Audio::Transcription, OpenAI::Models::Audio::TranscriptionDiarized, OpenAI::Models::Audio::TranscriptionVerbose]
         #
         # @see OpenAI::Models::Audio::TranscriptionCreateParams
         def create(params)
@@ -61,7 +65,7 @@ module OpenAI
         #
         # Transcribes audio into the input language.
         #
-        # @overload create_streaming(file:, model:, chunking_strategy: nil, include: nil, language: nil, prompt: nil, response_format: nil, temperature: nil, timestamp_granularities: nil, request_options: {})
+        # @overload create_streaming(file:, model:, chunking_strategy: nil, include: nil, known_speaker_names: nil, known_speaker_references: nil, language: nil, prompt: nil, response_format: nil, temperature: nil, timestamp_granularities: nil, request_options: {})
         #
         # @param file [Pathname, StringIO, IO, String, OpenAI::FilePart] The audio file object (not file name) to transcribe, in one of these formats: fl
         #
@@ -71,6 +75,10 @@ module OpenAI
         #
         # @param include [Array<Symbol, OpenAI::Models::Audio::TranscriptionInclude>] Additional information to include in the transcription response.
         #
+        # @param known_speaker_names [Array<String>] Optional list of speaker names that correspond to the audio samples provided in
+        #
+        # @param known_speaker_references [Array<String>] Optional list of audio samples (as [data URLs](https://developer.mozilla.org/en-
+        #
         # @param language [String] The language of the input audio. Supplying the input language in [ISO-639-1](htt
         #
         # @param prompt [String] An optional text to guide the model's style or continue a previous audio segment
@@ -83,7 +91,7 @@ module OpenAI
         #
         # @param request_options [OpenAI::RequestOptions, Hash{Symbol=>Object}, nil]
         #
-        # @return [OpenAI::Internal::Stream<OpenAI::Models::Audio::TranscriptionTextDeltaEvent, OpenAI::Models::Audio::TranscriptionTextDoneEvent>]
+        # @return [OpenAI::Internal::Stream<OpenAI::Models::Audio::TranscriptionTextSegmentEvent, OpenAI::Models::Audio::TranscriptionTextDeltaEvent, OpenAI::Models::Audio::TranscriptionTextDoneEvent>]
         #
         # @see OpenAI::Models::Audio::TranscriptionCreateParams
         def create_streaming(params)

data/lib/openai/resources/files.rb

@@ -105,7 +105,7 @@ module OpenAI
         )
       end
 
-      # Delete a file.
+      # Delete a file and remove it from all vector stores.
       #
       # @overload delete(file_id, request_options: {})
       #

data/lib/openai/resources/vector_stores.rb

@@ -14,10 +14,12 @@ module OpenAI
       #
       # Create a vector store.
       #
-      # @overload create(chunking_strategy: nil, expires_after: nil, file_ids: nil, metadata: nil, name: nil, request_options: {})
+      # @overload create(chunking_strategy: nil, description: nil, expires_after: nil, file_ids: nil, metadata: nil, name: nil, request_options: {})
       #
       # @param chunking_strategy [OpenAI::Models::AutoFileChunkingStrategyParam, OpenAI::Models::StaticFileChunkingStrategyObjectParam] The chunking strategy used to chunk the file(s). If not set, will use the `auto`
       #
+      # @param description [String] A description for the vector store. Can be used to describe the vector store's p
+      #
       # @param expires_after [OpenAI::Models::VectorStoreCreateParams::ExpiresAfter] The expiration policy for a vector store.
       #
       # @param file_ids [Array<String>] A list of [File](https://platform.openai.com/docs/api-reference/files) IDs that

data/lib/openai/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module OpenAI
-  VERSION = "0.30.0"
+  VERSION = "0.32.0"
 end

data/lib/openai.rb

@@ -79,11 +79,14 @@ require_relative "openai/models/audio/speech_model"
 require_relative "openai/models/audio/transcription"
 require_relative "openai/models/audio/transcription_create_params"
 require_relative "openai/models/audio/transcription_create_response"
+require_relative "openai/models/audio/transcription_diarized"
+require_relative "openai/models/audio/transcription_diarized_segment"
 require_relative "openai/models/audio/transcription_include"
 require_relative "openai/models/audio/transcription_segment"
 require_relative "openai/models/audio/transcription_stream_event"
 require_relative "openai/models/audio/transcription_text_delta_event"
 require_relative "openai/models/audio/transcription_text_done_event"
+require_relative "openai/models/audio/transcription_text_segment_event"
 require_relative "openai/models/audio/transcription_verbose"
 require_relative "openai/models/audio/transcription_word"
 require_relative "openai/models/audio/translation"

data/rbi/openai/models/audio/transcription_create_params.rbi

@@ -21,8 +21,8 @@ module OpenAI
         attr_accessor :file
 
         # 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).
+        # `gpt-4o-mini-transcribe`, `whisper-1` (which is powered by our open source
+        # Whisper V2 model), and `gpt-4o-transcribe-diarize`.
         sig { returns(T.any(String, OpenAI::AudioModel::OrSymbol)) }
         attr_accessor :model
 
@@ -30,6 +30,8 @@ module OpenAI
         # 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.
+        # Required when using `gpt-4o-transcribe-diarize` for inputs longer than 30
+        # seconds.
         sig do
           returns(
             T.nilable(
@@ -46,7 +48,8 @@ module OpenAI
         # 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`.
+        # `gpt-4o-mini-transcribe`. This field is not supported when using
+        # `gpt-4o-transcribe-diarize`.
         sig do
           returns(
             T.nilable(T::Array[OpenAI::Audio::TranscriptionInclude::OrSymbol])
@@ -61,6 +64,26 @@ module OpenAI
         end
         attr_writer :include
 
+        # Optional list of speaker names that correspond to the audio samples provided in
+        # `known_speaker_references[]`. Each entry should be a short identifier (for
+        # example `customer` or `agent`). Up to 4 speakers are supported.
+        sig { returns(T.nilable(T::Array[String])) }
+        attr_reader :known_speaker_names
+
+        sig { params(known_speaker_names: T::Array[String]).void }
+        attr_writer :known_speaker_names
+
+        # Optional list of audio samples (as
+        # [data URLs](https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/Data_URLs))
+        # that contain known speaker references matching `known_speaker_names[]`. Each
+        # sample must be between 2 and 10 seconds, and can use any of the same input audio
+        # formats supported by `file`.
+        sig { returns(T.nilable(T::Array[String])) }
+        attr_reader :known_speaker_references
+
+        sig { params(known_speaker_references: T::Array[String]).void }
+        attr_writer :known_speaker_references
+
         # 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.
@@ -73,7 +96,8 @@ module OpenAI
         # 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.
+        # should match the audio language. This field is not supported when using
+        # `gpt-4o-transcribe-diarize`.
         sig { returns(T.nilable(String)) }
         attr_reader :prompt
 
@@ -81,8 +105,10 @@ module OpenAI
         attr_writer :prompt
 
         # 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`.
+        # `verbose_json`, `vtt`, or `diarized_json`. For `gpt-4o-transcribe` and
+        # `gpt-4o-mini-transcribe`, the only supported format is `json`. For
+        # `gpt-4o-transcribe-diarize`, the supported formats are `json`, `text`, and
+        # `diarized_json`, with `diarized_json` required to receive speaker annotations.
         sig { returns(T.nilable(OpenAI::AudioResponseFormat::OrSymbol)) }
         attr_reader :response_format
 
@@ -106,7 +132,8 @@ module OpenAI
         # `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.
+        # incurs additional latency. This option is not available for
+        # `gpt-4o-transcribe-diarize`.
         sig do
           returns(
             T.nilable(
@@ -140,6 +167,8 @@ module OpenAI
                 )
               ),
             include: T::Array[OpenAI::Audio::TranscriptionInclude::OrSymbol],
+            known_speaker_names: T::Array[String],
+            known_speaker_references: T::Array[String],
             language: String,
             prompt: String,
             response_format: OpenAI::AudioResponseFormat::OrSymbol,
@@ -156,20 +185,33 @@ module OpenAI
           # flac, mp3, mp4, mpeg, mpga, m4a, ogg, wav, or webm.
           file:,
           # 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).
+          # `gpt-4o-mini-transcribe`, `whisper-1` (which is powered by our open source
+          # Whisper V2 model), and `gpt-4o-transcribe-diarize`.
           model:,
           # 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.
+          # Required when using `gpt-4o-transcribe-diarize` for inputs longer than 30
+          # seconds.
           chunking_strategy: nil,
           # 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`.
+          # `gpt-4o-mini-transcribe`. This field is not supported when using
+          # `gpt-4o-transcribe-diarize`.
           include: nil,
+          # Optional list of speaker names that correspond to the audio samples provided in
+          # `known_speaker_references[]`. Each entry should be a short identifier (for
+          # example `customer` or `agent`). Up to 4 speakers are supported.
+          known_speaker_names: nil,
+          # Optional list of audio samples (as
+          # [data URLs](https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/Data_URLs))
+          # that contain known speaker references matching `known_speaker_names[]`. Each
+          # sample must be between 2 and 10 seconds, and can use any of the same input audio
+          # formats supported by `file`.
+          known_speaker_references: nil,
           # 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.
@@ -177,11 +219,14 @@ module OpenAI
           # 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.
+          # should match the audio language. This field is not supported when using
+          # `gpt-4o-transcribe-diarize`.
           prompt: nil,
           # 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`.
+          # `verbose_json`, `vtt`, or `diarized_json`. For `gpt-4o-transcribe` and
+          # `gpt-4o-mini-transcribe`, the only supported format is `json`. For
+          # `gpt-4o-transcribe-diarize`, the supported formats are `json`, `text`, and
+          # `diarized_json`, with `diarized_json` required to receive speaker annotations.
           response_format: nil,
           # 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
@@ -193,7 +238,8 @@ module OpenAI
           # `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.
+          # incurs additional latency. This option is not available for
+          # `gpt-4o-transcribe-diarize`.
           timestamp_granularities: nil,
           request_options: {}
         )
@@ -212,6 +258,8 @@ module OpenAI
                   )
                 ),
               include: T::Array[OpenAI::Audio::TranscriptionInclude::OrSymbol],
+              known_speaker_names: T::Array[String],
+              known_speaker_references: T::Array[String],
               language: String,
               prompt: String,
               response_format: OpenAI::AudioResponseFormat::OrSymbol,
@@ -228,8 +276,8 @@ module OpenAI
         end
 
         # 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).
+        # `gpt-4o-mini-transcribe`, `whisper-1` (which is powered by our open source
+        # Whisper V2 model), and `gpt-4o-transcribe-diarize`.
         module Model
           extend OpenAI::Internal::Type::Union
 
@@ -251,6 +299,8 @@ module OpenAI
         # 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.
+        # Required when using `gpt-4o-transcribe-diarize` for inputs longer than 30
+        # seconds.
         module ChunkingStrategy
           extend OpenAI::Internal::Type::Union
 

data/rbi/openai/models/audio/transcription_create_response.rbi

@@ -12,6 +12,7 @@ module OpenAI
           T.type_alias do
             T.any(
               OpenAI::Audio::Transcription,
+              OpenAI::Audio::TranscriptionDiarized,
               OpenAI::Audio::TranscriptionVerbose
             )
           end