All AI models · LLM benchmarks · Methodology

OpenAI

GPT-Realtime-2.1

Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior

Model facts

Context window
128K tokens
Maximum output
32K tokens
Cheapest paid input
$4 per 1M tokens
Cheapest paid output
$24 per 1M tokens
Open weights
no
Providers
2

Capabilities and modalities

reasoning, tool calling, attachments, temperature control, text, image, audio.

Observed price history

  • 2026-08-24: $4 input / $24 output per 1M tokens

API providers