🎙️OpenAI Ships GPT-Realtime-2 With GPT-5 Reasoning and 128K Voice Context
OpenAI Ships GPT-Realtime-2 With GPT-5 Reasoning and 128K V…
TL;DR
OpenAI launched three new Realtime API models — GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper — with GPT-5-class reasoning, a 128K context w…
OpenAI launched three new Realtime API models — GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper — with GPT-5-class reasoning, a 128K context window (up from 32K), parallel tool calls, and 'preamble' phrases that let agents say 'one moment' while they work.
Key Points
Context window jumps from 32K to 128K — long voice agent flows now feasible
Parallel tool calls and preamble phrases for natural multi-step responses
Pricing: $32/M audio input, $64/M audio output; Translate at $0.034/min
Why It Matters
Voice agents have been the demo that doesn't ship. A 128K context plus parallel tool calls is finally enough to build call-center-grade flows that don't feel like 2024 IVR.
Frequently Asked Questions
Why does this matter?
Voice agents have been the demo that doesn't ship. A 128K context plus parallel tool calls is finally enough to build call-center-grade flows that don't feel like 2024 IVR.
What happened?
OpenAI launched three new Realtime API models — GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper — with GPT-5-class reasoning, a 128K context w…
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,303 builders reading daily.