GPT Audio Mini
Neural Network
GPT Audio Mini is OpenAI's cost-effective audio model for natural speech synthesis and voice scenarios.
Specifications
- Input
- text, documents
- Output
- text
- Context size
- 128 000 tokens
- Max answer length
- 16 384 tokens
- Capabilities
- function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works GPT Audio Mini?
GPT Audio Mini is a lightweight version of OpenAI's GPT Audio, designed for cost-effective speech processing. The new snapshot features an updated decoder: the voice sounds more natural, and the timbre remains stable throughout the utterance—this is how the model differs from the previous snapshot. Through BotHub, you connect without a VPN or foreign card, pay in rubles only for tokens used rather than a subscription, and switch between hundreds of models in one window if a task requires another. A unified OpenAI-compatible API allows you to integrate voiceover into your product and replace the model without rewriting the integration, while data is protected by AES-GCM encryption. It is useful for voicing training courses or podcast episodes, building voice assistants for support, recording character lines for games or prototypes, or adding audio to presentations. Teams have access to contracts, invoices, electronic document management (EDM), and an admin panel with user limits.Frequently asked questions about GPT Audio Mini
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
OpenAI's gpt-audio-mini works with text and documents: it answers questions, summarizes and analyzes uploaded files, and helps with code and drafts. The model supports function calling, making it suitable for scenarios where the response needs to be integrated into your product via API.
The model generates up to 16,384 tokens per response—enough for a long article, detailed instruction, or substantial code snippet. The context window is 128,000 tokens, so you can include a long document and the entire conversation history in your request.
A separate reasoning mode is not explicitly noted in our data, but that doesn't mean it doesn't exist. In practice, a simple trick helps: ask in the prompt to break down the task step-by-step and then provide a final conclusion—this noticeably helps with logic and calculations.
Yes, both are confirmed by our data. The model can call functions and return structured responses according to a specified JSON schema, which is convenient for agents and automation. Through the unified OpenAI-compatible BotHub API, this can be integrated into your code without rewriting the integration.
According to our data, it accepts text and documents as input: you upload a file and ask to analyze its content. We have no notes on it accepting images, audio, or video, so for such tasks, it is more convenient to switch to a multimodal model in the same window.
We do not have data on prompt and response languages, so we cannot make specific promises. The easiest way to check is with your own task: run a couple of typical requests in BotHub and compare with another model. Switching takes seconds; no VPN or foreign card is needed.