GPT-Realtime-2
OpenAI’s most intelligent voice model yet.
An OpenAI API real-time voice model for live conversational agents, described as adding GPT-5-class reasoning to voice interactions.
Pricing
Official site · Nov 1, 2026, 6:29 AM
Input / 1M
$4.00
Output / 1M
$24.00
Cached input / 1M
$0.40
Per 1M tokens: Text input $4, cached input $0.40, output $24; audio input $32, cached input $0.40, output $64; image input $5, cached input $0.50 (image output pricing not listed).
OpenAI's official model documentation explicitly lists token pricing across text, audio, and image inputs, confirming multimodal pricing.
Model Intelligence
Benchmarkable
Yes
Model level
release
Recent stories
0 linked stories
No linked stories yet.