GPT-4o Audio Models by OpenAI
OpenAIProprietary
OpenAI releases · 44 in the last 12 monthsChangelog →
2 models2024Up to 128K ctx
Last refreshed 2026-10-17. Next refresh: weekly.
Capabilities
VisionAll models
Code ExecutionAll models
Links
WebsiteAbout
GPT-4o Audio is OpenAI’s GPT-4o line for models that take and produce audio as well as text. It sits on GPT-4o (“o” for omni), OpenAI’s flagship trained end-to-end across text, vision, and audio so one network can hear tone and speakers and reply with speech, not only a separate speech-to-text then text then text-to-speech pipeline. In the API, the GPT-4o Audio preview models accept text and audio input and return text and audio output through Chat Completions.
Decision facts
- Best fit
- audiovision and multimodal workcode execution
- Capability starting point
- GPT-4o Audio Preview (12-17) with 128K context and multimodal inputs
- Lowest tracked input
- Not tracked
- Closest related family
- GPT-5
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
1 in view1 retired
GPT-4o Audio Preview (12-17)Current
Use when the workload needs audio, 128K context, and code execution.
2024-12audio128K contextcode execution
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| GPT-4o Audio Preview (12-17) | Use when the workload needs audio, 128K context, and code execution. | 2024-12 | audio128K contextcode execution | Current |
Release Timeline
2 release groups2024-12
1 current
GPT-4o Audio Preview (12-17)
Currentaudio128K contextcode execution
2024-10
1 retired
GPT-4o Audio Preview (10-01)
Archivedaudio128K contextcode execution
Specifications(2 models)
| Model | Released | Context | Vision | Code Exec |
|---|---|---|---|---|
| GPT-4o Audio Preview (12-17) | 2024-12 | 128K | Yes | Yes |
