openai 0.22.1 → 0.23.0

data/sig/openai/models/realtime/realtime_tools_config_union.rbs

@@ -2,64 +2,12 @@ module OpenAI
   module Models
     module Realtime
       type realtime_tools_config_union =
-        OpenAI::Realtime::RealtimeToolsConfigUnion::Function
+        OpenAI::Realtime::Models
         | OpenAI::Realtime::RealtimeToolsConfigUnion::Mcp
 
       module RealtimeToolsConfigUnion
         extend OpenAI::Internal::Type::Union
 
-        type function =
-          {
-            description: String,
-            name: String,
-            parameters: top,
-            type: OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_
-          }
-
-        class Function < OpenAI::Internal::Type::BaseModel
-          attr_reader description: String?
-
-          def description=: (String) -> String
-
-          attr_reader name: String?
-
-          def name=: (String) -> String
-
-          attr_reader parameters: top?
-
-          def parameters=: (top) -> top
-
-          attr_reader type: OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_?
-
-          def type=: (
-            OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_
-          ) -> OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_
-
-          def initialize: (
-            ?description: String,
-            ?name: String,
-            ?parameters: top,
-            ?type: OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_
-          ) -> void
-
-          def to_hash: -> {
-            description: String,
-            name: String,
-            parameters: top,
-            type: OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_
-          }
-
-          type type_ = :function
-
-          module Type
-            extend OpenAI::Internal::Type::Enum
-
-            FUNCTION: :function
-
-            def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeToolsConfigUnion::Function::type_]
-          end
-        end
-
         type mcp =
           {
             server_label: String,

data/sig/openai/models/realtime/realtime_transcription_session_audio.rbs

@@ -0,0 +1,24 @@
+module OpenAI
+  module Models
+    module Realtime
+      type realtime_transcription_session_audio =
+        { input: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput }
+
+      class RealtimeTranscriptionSessionAudio < OpenAI::Internal::Type::BaseModel
+        attr_reader input: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput?
+
+        def input=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput
+
+        def initialize: (
+          ?input: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput
+        ) -> void
+
+        def to_hash: -> {
+          input: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput
+        }
+      end
+    end
+  end
+end

data/sig/openai/models/realtime/realtime_transcription_session_audio_input.rbs

@@ -0,0 +1,72 @@
+module OpenAI
+  module Models
+    module Realtime
+      type realtime_transcription_session_audio_input =
+        {
+          format_: OpenAI::Models::Realtime::realtime_audio_formats,
+          noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction,
+          transcription: OpenAI::Realtime::AudioTranscription,
+          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection
+        }
+
+      class RealtimeTranscriptionSessionAudioInput < OpenAI::Internal::Type::BaseModel
+        attr_reader format_: OpenAI::Models::Realtime::realtime_audio_formats?
+
+        def format_=: (
+          OpenAI::Models::Realtime::realtime_audio_formats
+        ) -> OpenAI::Models::Realtime::realtime_audio_formats
+
+        attr_reader noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction?
+
+        def noise_reduction=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction
+
+        attr_reader transcription: OpenAI::Realtime::AudioTranscription?
+
+        def transcription=: (
+          OpenAI::Realtime::AudioTranscription
+        ) -> OpenAI::Realtime::AudioTranscription
+
+        attr_reader turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection?
+
+        def turn_detection=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection
+
+        def initialize: (
+          ?format_: OpenAI::Models::Realtime::realtime_audio_formats,
+          ?noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction,
+          ?transcription: OpenAI::Realtime::AudioTranscription,
+          ?turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection
+        ) -> void
+
+        def to_hash: -> {
+          format_: OpenAI::Models::Realtime::realtime_audio_formats,
+          noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInput::NoiseReduction,
+          transcription: OpenAI::Realtime::AudioTranscription,
+          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection
+        }
+
+        type noise_reduction =
+          { type: OpenAI::Models::Realtime::noise_reduction_type }
+
+        class NoiseReduction < OpenAI::Internal::Type::BaseModel
+          attr_reader type: OpenAI::Models::Realtime::noise_reduction_type?
+
+          def type=: (
+            OpenAI::Models::Realtime::noise_reduction_type
+          ) -> OpenAI::Models::Realtime::noise_reduction_type
+
+          def initialize: (
+            ?type: OpenAI::Models::Realtime::noise_reduction_type
+          ) -> void
+
+          def to_hash: -> {
+            type: OpenAI::Models::Realtime::noise_reduction_type
+          }
+        end
+      end
+    end
+  end
+end

data/sig/openai/models/realtime/realtime_transcription_session_audio_input_turn_detection.rbs

@@ -0,0 +1,99 @@
+module OpenAI
+  module Models
+    module Realtime
+      type realtime_transcription_session_audio_input_turn_detection =
+        {
+          create_response: bool,
+          eagerness: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness,
+          idle_timeout_ms: Integer?,
+          interrupt_response: bool,
+          prefix_padding_ms: Integer,
+          silence_duration_ms: Integer,
+          threshold: Float,
+          type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_
+        }
+
+      class RealtimeTranscriptionSessionAudioInputTurnDetection < OpenAI::Internal::Type::BaseModel
+        attr_reader create_response: bool?
+
+        def create_response=: (bool) -> bool
+
+        attr_reader eagerness: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness?
+
+        def eagerness=: (
+          OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness
+        ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness
+
+        attr_accessor idle_timeout_ms: Integer?
+
+        attr_reader interrupt_response: bool?
+
+        def interrupt_response=: (bool) -> bool
+
+        attr_reader prefix_padding_ms: Integer?
+
+        def prefix_padding_ms=: (Integer) -> Integer
+
+        attr_reader silence_duration_ms: Integer?
+
+        def silence_duration_ms=: (Integer) -> Integer
+
+        attr_reader threshold: Float?
+
+        def threshold=: (Float) -> Float
+
+        attr_reader type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_?
+
+        def type=: (
+          OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_
+        ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_
+
+        def initialize: (
+          ?create_response: bool,
+          ?eagerness: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness,
+          ?idle_timeout_ms: Integer?,
+          ?interrupt_response: bool,
+          ?prefix_padding_ms: Integer,
+          ?silence_duration_ms: Integer,
+          ?threshold: Float,
+          ?type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_
+        ) -> void
+
+        def to_hash: -> {
+          create_response: bool,
+          eagerness: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness,
+          idle_timeout_ms: Integer?,
+          interrupt_response: bool,
+          prefix_padding_ms: Integer,
+          silence_duration_ms: Integer,
+          threshold: Float,
+          type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_
+        }
+
+        type eagerness = :low | :medium | :high | :auto
+
+        module Eagerness
+          extend OpenAI::Internal::Type::Enum
+
+          LOW: :low
+          MEDIUM: :medium
+          HIGH: :high
+          AUTO: :auto
+
+          def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::eagerness]
+        end
+
+        type type_ = :server_vad | :semantic_vad
+
+        module Type
+          extend OpenAI::Internal::Type::Enum
+
+          SERVER_VAD: :server_vad
+          SEMANTIC_VAD: :semantic_vad
+
+          def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionAudioInputTurnDetection::type_]
+        end
+      end
+    end
+  end
+end

data/sig/openai/models/realtime/realtime_transcription_session_client_secret.rbs

@@ -0,0 +1,20 @@
+module OpenAI
+  module Models
+    class RealtimeTranscriptionSessionClientSecret = Realtime::RealtimeTranscriptionSessionClientSecret
+
+    module Realtime
+      type realtime_transcription_session_client_secret =
+        { expires_at: Integer, value: String }
+
+      class RealtimeTranscriptionSessionClientSecret < OpenAI::Internal::Type::BaseModel
+        attr_accessor expires_at: Integer
+
+        attr_accessor value: String
+
+        def initialize: (expires_at: Integer, value: String) -> void
+
+        def to_hash: -> { expires_at: Integer, value: String }
+      end
+    end
+  end
+end

data/sig/openai/models/realtime/realtime_transcription_session_create_request.rbs

@@ -3,86 +3,38 @@ module OpenAI
     module Realtime
       type realtime_transcription_session_create_request =
         {
-          model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::model,
           type: :transcription,
-          include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_],
-          input_audio_format: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format,
-          input_audio_noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction,
-          input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription,
-          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection
+          audio: OpenAI::Realtime::RealtimeTranscriptionSessionAudio,
+          include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]
         }
 
       class RealtimeTranscriptionSessionCreateRequest < OpenAI::Internal::Type::BaseModel
-        attr_accessor model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::model
-
         attr_accessor type: :transcription
 
+        attr_reader audio: OpenAI::Realtime::RealtimeTranscriptionSessionAudio?
+
+        def audio=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionAudio
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionAudio
+
         attr_reader include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]?
 
         def include=: (
           ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]
         ) -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]
 
-        attr_reader input_audio_format: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format?
-
-        def input_audio_format=: (
-          OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format
-        ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format
-
-        attr_reader input_audio_noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction?
-
-        def input_audio_noise_reduction=: (
-          OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction
-        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction
-
-        attr_reader input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription?
-
-        def input_audio_transcription=: (
-          OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription
-        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription
-
-        attr_reader turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection?
-
-        def turn_detection=: (
-          OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection
-        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection
-
         def initialize: (
-          model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::model,
+          ?audio: OpenAI::Realtime::RealtimeTranscriptionSessionAudio,
           ?include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_],
-          ?input_audio_format: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format,
-          ?input_audio_noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction,
-          ?input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription,
-          ?turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection,
           ?type: :transcription
         ) -> void
 
         def to_hash: -> {
-          model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::model,
           type: :transcription,
-          include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_],
-          input_audio_format: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format,
-          input_audio_noise_reduction: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction,
-          input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription,
-          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection
+          audio: OpenAI::Realtime::RealtimeTranscriptionSessionAudio,
+          include: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]
         }
 
-        type model =
-          String
-          | :"whisper-1"
-          | :"gpt-4o-transcribe"
-          | :"gpt-4o-mini-transcribe"
-
-        module Model
-          extend OpenAI::Internal::Type::Union
-
-          def self?.variants: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::model]
-
-          WHISPER_1: :"whisper-1"
-          GPT_4O_TRANSCRIBE: :"gpt-4o-transcribe"
-          GPT_4O_MINI_TRANSCRIBE: :"gpt-4o-mini-transcribe"
-        end
-
         type include_ = :"item.input_audio_transcription.logprobs"
 
         module Include
@@ -92,150 +44,6 @@ module OpenAI
 
           def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::include_]
         end
-
-        type input_audio_format = :pcm16 | :g711_ulaw | :g711_alaw
-
-        module InputAudioFormat
-          extend OpenAI::Internal::Type::Enum
-
-          PCM16: :pcm16
-          G711_ULAW: :g711_ulaw
-          G711_ALAW: :g711_alaw
-
-          def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::input_audio_format]
-        end
-
-        type input_audio_noise_reduction =
-          {
-            type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_
-          }
-
-        class InputAudioNoiseReduction < OpenAI::Internal::Type::BaseModel
-          attr_reader type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_?
-
-          def type=: (
-            OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_
-          ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_
-
-          def initialize: (
-            ?type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_
-          ) -> void
-
-          def to_hash: -> {
-            type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_
-          }
-
-          type type_ = :near_field | :far_field
-
-          module Type
-            extend OpenAI::Internal::Type::Enum
-
-            NEAR_FIELD: :near_field
-            FAR_FIELD: :far_field
-
-            def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioNoiseReduction::type_]
-          end
-        end
-
-        type input_audio_transcription =
-          {
-            language: String,
-            model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model,
-            prompt: String
-          }
-
-        class InputAudioTranscription < OpenAI::Internal::Type::BaseModel
-          attr_reader language: String?
-
-          def language=: (String) -> String
-
-          attr_reader model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model?
-
-          def model=: (
-            OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model
-          ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model
-
-          attr_reader prompt: String?
-
-          def prompt=: (String) -> String
-
-          def initialize: (
-            ?language: String,
-            ?model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model,
-            ?prompt: String
-          ) -> void
-
-          def to_hash: -> {
-            language: String,
-            model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model,
-            prompt: String
-          }
-
-          type model =
-            :"gpt-4o-transcribe" | :"gpt-4o-mini-transcribe" | :"whisper-1"
-
-          module Model
-            extend OpenAI::Internal::Type::Enum
-
-            GPT_4O_TRANSCRIBE: :"gpt-4o-transcribe"
-            GPT_4O_MINI_TRANSCRIBE: :"gpt-4o-mini-transcribe"
-            WHISPER_1: :"whisper-1"
-
-            def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::InputAudioTranscription::model]
-          end
-        end
-
-        type turn_detection =
-          {
-            prefix_padding_ms: Integer,
-            silence_duration_ms: Integer,
-            threshold: Float,
-            type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_
-          }
-
-        class TurnDetection < OpenAI::Internal::Type::BaseModel
-          attr_reader prefix_padding_ms: Integer?
-
-          def prefix_padding_ms=: (Integer) -> Integer
-
-          attr_reader silence_duration_ms: Integer?
-
-          def silence_duration_ms=: (Integer) -> Integer
-
-          attr_reader threshold: Float?
-
-          def threshold=: (Float) -> Float
-
-          attr_reader type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_?
-
-          def type=: (
-            OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_
-          ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_
-
-          def initialize: (
-            ?prefix_padding_ms: Integer,
-            ?silence_duration_ms: Integer,
-            ?threshold: Float,
-            ?type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_
-          ) -> void
-
-          def to_hash: -> {
-            prefix_padding_ms: Integer,
-            silence_duration_ms: Integer,
-            threshold: Float,
-            type: OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_
-          }
-
-          type type_ = :server_vad
-
-          module Type
-            extend OpenAI::Internal::Type::Enum
-
-            SERVER_VAD: :server_vad
-
-            def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateRequest::TurnDetection::type_]
-          end
-        end
       end
     end
   end

data/sig/openai/models/realtime/realtime_transcription_session_create_response.rbs

@@ -0,0 +1,69 @@
+module OpenAI
+  module Models
+    class RealtimeTranscriptionSessionCreateResponse = Realtime::RealtimeTranscriptionSessionCreateResponse
+
+    module Realtime
+      type realtime_transcription_session_create_response =
+        {
+          client_secret: OpenAI::Realtime::RealtimeTranscriptionSessionClientSecret,
+          input_audio_format: String,
+          input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription,
+          modalities: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality],
+          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection
+        }
+
+      class RealtimeTranscriptionSessionCreateResponse < OpenAI::Internal::Type::BaseModel
+        attr_accessor client_secret: OpenAI::Realtime::RealtimeTranscriptionSessionClientSecret
+
+        attr_reader input_audio_format: String?
+
+        def input_audio_format=: (String) -> String
+
+        attr_reader input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription?
+
+        def input_audio_transcription=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription
+
+        attr_reader modalities: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality]?
+
+        def modalities=: (
+          ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality]
+        ) -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality]
+
+        attr_reader turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection?
+
+        def turn_detection=: (
+          OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection
+        ) -> OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection
+
+        def initialize: (
+          client_secret: OpenAI::Realtime::RealtimeTranscriptionSessionClientSecret,
+          ?input_audio_format: String,
+          ?input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription,
+          ?modalities: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality],
+          ?turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection
+        ) -> void
+
+        def to_hash: -> {
+          client_secret: OpenAI::Realtime::RealtimeTranscriptionSessionClientSecret,
+          input_audio_format: String,
+          input_audio_transcription: OpenAI::Realtime::RealtimeTranscriptionSessionInputAudioTranscription,
+          modalities: ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality],
+          turn_detection: OpenAI::Realtime::RealtimeTranscriptionSessionTurnDetection
+        }
+
+        type modality = :text | :audio
+
+        module Modality
+          extend OpenAI::Internal::Type::Enum
+
+          TEXT: :text
+          AUDIO: :audio
+
+          def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionCreateResponse::modality]
+        end
+      end
+    end
+  end
+end

data/sig/openai/models/realtime/realtime_transcription_session_input_audio_transcription.rbs

@@ -0,0 +1,59 @@
+module OpenAI
+  module Models
+    class RealtimeTranscriptionSessionInputAudioTranscription = Realtime::RealtimeTranscriptionSessionInputAudioTranscription
+
+    module Realtime
+      type realtime_transcription_session_input_audio_transcription =
+        {
+          language: String,
+          model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model,
+          prompt: String
+        }
+
+      class RealtimeTranscriptionSessionInputAudioTranscription < OpenAI::Internal::Type::BaseModel
+        attr_reader language: String?
+
+        def language=: (String) -> String
+
+        attr_reader model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model?
+
+        def model=: (
+          OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model
+        ) -> OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model
+
+        attr_reader prompt: String?
+
+        def prompt=: (String) -> String
+
+        def initialize: (
+          ?language: String,
+          ?model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model,
+          ?prompt: String
+        ) -> void
+
+        def to_hash: -> {
+          language: String,
+          model: OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model,
+          prompt: String
+        }
+
+        type model =
+          :"whisper-1"
+          | :"gpt-4o-transcribe-latest"
+          | :"gpt-4o-mini-transcribe"
+          | :"gpt-4o-transcribe"
+
+        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"
+
+          def self?.values: -> ::Array[OpenAI::Models::Realtime::RealtimeTranscriptionSessionInputAudioTranscription::model]
+        end
+      end
+    end
+  end
+end