Very long inputs
Choose GPT-Audio Mini
GPT-Audio Mini reads up to 128k tokens per request, against 16k for GPT-Realtime-Translate.
Choose GPT-Audio Mini for very long inputs. Choose GPT-Realtime-Translate for the latest release and the most recent knowledge.
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Choose GPT-Audio Mini
GPT-Audio Mini reads up to 128k tokens per request, against 16k for GPT-Realtime-Translate.
Choose GPT-Realtime-Translate
GPT-Realtime-Translate is the newer release (May 7, 2026, against Oct 6, 2025 for GPT-Audio Mini).
Choose GPT-Realtime-Translate
GPT-Realtime-Translate was trained on data up to September 2024, against October 2023 for GPT-Audio Mini.
| GPT-Audio Mini | GPT-Realtime-Translate | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Category | Voice and realtime | Voice and realtime |
| Status | Deprecated | Generally available |
| Headline price | No data | $0.034 / minute |
| Context window | 128k | 16k |
| Max output | 16k | 2k |
| Released | Oct 6, 2025 | May 7, 2026 |
| Knowledge cutoff | October 2023 | September 2024 |
| Licence | Proprietary | Proprietary |
| Capabilities | Audio input | Audio input |
| API model id | gpt-audio-mini | gpt-realtime-translate |
Choose GPT-Audio Mini for very long inputs. Choose GPT-Realtime-Translate for the latest release and the most recent knowledge.
A price comparison is not possible: one of them has no comparable published price.
GPT-Audio Mini reads more per request: 128k tokens against 16k.
GPT-Audio Mini is made by OpenAI and GPT-Realtime-Translate by OpenAI. A price comparison is not possible: one of them has no comparable published price. GPT-Audio Mini reads more per request: 128k tokens against 16k. GPT-Realtime-Translate is the newer model (May 7, 2026, against Oct 6, 2025).
GPT-Realtime-Translate was released on May 7, 2026 and GPT-Audio Mini on Oct 6, 2025.