OpenAI
GPT Audio
Released 2026-01-19
The gpt-audio model is OpenAI's first generally available audio model. The new snapshot features an upgraded decoder for more natural sounding voices and maintains better voice consistency. Audio is priced...
You send / 1M
$2.50
tokens in
It replies / 1M
$10.00
tokens out
Memory
128K
tokens it can read at once
Typical request
$0.0083
1,000 words in, 400 out
Data privacy
No training or retention: Unknown. Check the provider's terms before sending sensitive data.
Based on OpenRouter and Vercel AI Gateway data. Confirm terms with the provider before sending sensitive data.
What it can do
- Reads: text, audio
- Writes: text, audio
- Reasoning: No
- Tool calling: Yes
- Longest reply: 16K tokens
Similar models
- Command ACohere · $2.50 in / $10.00 out per 1M
- Command R+ (08-2024)Cohere · $2.50 in / $10.00 out per 1M
- Palmyra X5Writer · $0.600 in / $6.00 out per 1M
- Mistral Large 2407Mistral · $2.00 in / $6.00 out per 1M
- Mixtral 8x22B InstructMistral · $2.00 in / $6.00 out per 1M
- Mistral LargeMistral · $2.00 in / $6.00 out per 1M