Sponsored by Byond Boundrys Consulting - Empowering Ideas, Delivering Results
OpenAI Intermediate

GPT-Audio-1.5

GPT-Audio-1.5 is OpenAI’s audio-in, audio-out model for voice assistants and spoken apps with tool calling, a 128K context and $32/M audio input pricing.

Speech & Voice ModelsTextAudio Paid
In plain English

What is this model and why does it matter?

GPT-Audio-1.5 is OpenAI's general-availability audio-in/audio-out model for Chat Completions, supporting spoken interactions, audio understanding and tool-enabled conversational workflows.

Voice assistantsAudio understandingSpoken interfacesConversational applicationsFunction-enabled audio workflows
Model overview

GPT-Audio-1.5: features, use cases and important details

GPT-Audio-1.5 is a current AI model verified from first-party OpenAI sources.

GPT-Audio-1.5 verified specifications

GPT-Audio-1.5 is OpenAI’s general-availability audio-in/audio-out model for Chat Completions, supporting spoken interactions, audio understanding and tool-enabled conversational workflows. Its verified context or usage limit is 128000 tokens, with 16384 tokens maximum output.

GPT-Audio-1.5 pricing and access

GPT-Audio-1.5 costs $2.50 per 1M text input tokens and $10 per 1M text output tokens. Audio costs $32 per 1M input audio tokens and $64 per 1M output audio tokens.

GPT-Audio-1.5 best uses

Voice assistants, Audio understanding, Spoken interfaces, Conversational applications, Function-enabled audio workflows.

GPT-Audio-1.5 limitations

No image or video input, Audio cost can dominate high-volume workloads, Voice workflows still need latency testing.

Get started

How to use this model

  1. Create an OpenAI API key.
  2. Call gpt-audio-1.5 through Chat Completions.
  3. Provide text and/or audio input.
  4. Request text or audio output.
  5. Add function tools where needed.
  6. Track audio-token usage separately from text.
Copy and try

Example prompts

  • Listen to this audio and answer the user's question.
  • Respond to the user naturally with spoken audio.
  • Use the available function after understanding the spoken request.
Capabilities

What it can do

  • 128K context
  • 16K output
  • Audio input/output
  • Text input/output
  • Function calling
  • Streaming
Best for

Practical use cases

  • Voice bots
  • Audio Q&A
  • Accessibility
  • Spoken customer support
  • Interactive assistants
Pricing

What does it cost?

GPT-Audio-1.5 costs $2.50 per 1M text input tokens and $10 per 1M text output tokens. Audio costs $32 per 1M input audio tokens and $64 per 1M output audio tokens.

Input$2.50/M text; $32/M audio
Output$10/M text; $64/M audio
Simple summaryText and audio use separate token rates, with audio substantially more expensive than text.

What stands out

  • GA audio model
  • Audio in and out
  • Function calling
  • Large context

Things to consider

  • Audio tokens are expensive
  • No structured outputs
  • No fine-tuning
Limitations

Important restrictions and trade-offs

  • No image or video input
  • Audio cost can dominate high-volume workloads
  • Voice workflows still need latency testing
SimplifyAITools verdict

Our editorial take

A strong current OpenAI audio model for developers who need flexible audio input and output without using the Realtime-specific model family.

References

Primary sources

  1. Open source 1 ↗
  2. Open source 2 ↗
Keep exploring

Related AI models