---
last_updated: 2026-11-21
source_url: https://opper.ai/openai/gpt-realtime-2
title: "GPT Realtime 2 by OpenAI"
description: "GPT Realtime 2 by OpenAI: realtime voice model with GPT-5-class reasoning, vision input and configurable reasoning effort."
---

# GPT Realtime 2

By [OpenAI](https://opper.ai/openai).

Released 7 May 2026 · [Announcement](https://developers.openai.com/api/docs/changelog) · [All OpenAI releases](https://opper.ai/model-releases/openai)

GPT Realtime 2, released in May 2026, is OpenAI's first voice model with GPT-5-class reasoning, built for complex voice-agent workflows that need deeper reasoning and tool coordination. It accepts text, audio, and image inputs and produces text and audio outputs.

The model operates with a 128,000 token context window and up to 32,000 output tokens, with configurable reasoning effort levels (minimal, low, medium, high, xhigh). Against GPT Realtime 1.5 it lifts Big Bench Audio from 81.4 to 96.6 and Audio MultiChallenge from 34.7 to 48.5.

It handles interruptions, corrections, and parallel tool execution while keeping conversation natural, and can pause to reason before responding, making it well suited to agentic voice applications.

## Available routes

GPT Realtime 2 runs on 1 route through the Opper gateway. Prices are per 1M tokens. Its routes and prices, live from the catalogue, are public JSON with no API key needed at https://api.opper.ai/v3/models?q=openai%2Fgpt-realtime-2.

| Provider | Model id | Region | Zero data retention | Training | Input | Output | Cache read |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [OpenAI](https://opper.ai/provider/openai) | `openai/gpt-realtime-2` | US (United States) | Abuse monitoring | No | $4.00 | $24.00 | $0.40 |

## Key info

- Model id: `openai/gpt-realtime-2`
- Input: Text, Image, Audio
- Output: Text, Audio
- Features: Reasoning, Tools
- Context window: 128K
- Max output: 32K
- Text input price: $4.00 per 1M tokens ($0.40 cache read)
- Text output price: $24.00 per 1M tokens
- Image input price: $5.00 per 1M tokens
- Audio input price: $32.00 per 1M tokens
- Audio output price: $64.00 per 1M tokens
- Released: 7 May 2026
- US residency available
- Zero data retention with abuse monitoring
- No training by default

## Data handling per route

Each route hosting GPT Realtime 2 has its own privacy posture, residency, and GDPR terms. Postures are maintained by Opper with a last-verification timestamp.

### OpenAI — United States

Zero data retention with abuse monitoring: prompts and outputs are kept for up to 30 days for abuse monitoring only, and are not used for training. No training on customer data. US; SCCs.

- Zero data retention: With abuse monitoring, 30-day window.
- Training: No training on customer data.
- Logging: Abuse monitoring (30-day retention)
- Abuse monitoring: On by default, holds flagged content, 30 days, a person can view samples
- Caching: Content cached for replay
- Transfer mechanism: SCCs

Verified by Opper on 2026-09-16.

## Get started

Call GPT Realtime 2 through the Opper gateway with one API key. Realtime models run over a WebSocket at wss://api.opper.ai/v3/realtime: open it, send session.start with this model, then stream audio both ways.

See the [Opper realtime quickstart](https://docs.opper.ai/build/realtime/quickstart), the [realtime protocol](https://docs.opper.ai/build/realtime/protocol) and the [Opper multimodal skill](https://skills.opper.ai/opper-multimodal/SKILL.md).

GPT Realtime 2 is a premium model on Opper. Sign up needs no credit card: you get an API key straight away and the free models work in the playground and the API. Add a card to use premium models, pay-as-you-go with no minimum.

### Set it up with your agent

Paste this into a coding agent like Claude Code, Cursor or Codex and it will wire up Opper for you:

```text
Use curl to download, read and follow: https://skills.opper.ai/opper-multimodal/SKILL.md Start a realtime voice session with GPT Realtime 2
```

### Or call it directly

Node.js:

```javascript
import WebSocket from "ws";

const ws = new WebSocket("wss://api.opper.ai/v3/realtime", {
  headers: { Authorization: `Bearer ${process.env.OPPER_API_KEY}` },
});

ws.on("open", () => {
  ws.send(JSON.stringify({
    type: "session.start",
    config: {
      model: "openai/gpt-realtime-2",
      voice: "marin",
      instructions: "You are a concise voice assistant.",
    },
  }));
});

// Stream microphone audio in with { type: "audio.append", audio: <base64 PCM16> }
ws.on("message", (raw) => {
  const ev = JSON.parse(raw.toString());
  if (ev.type === "audio.delta") {
    // ev.audio is base64 PCM16: play it
  }
});
```

## Use GPT Realtime 2 in your apps

Pick your tool to see how to run it on GPT Realtime 2 through Opper.

- [Claude Code](https://opper.ai/apps/claude-code/gpt-realtime-2)
- [Claude Cowork](https://opper.ai/apps/claude-cowork/gpt-realtime-2)
- [Codex](https://opper.ai/apps/codex/gpt-realtime-2)
- [DeepSeek Harness](https://opper.ai/apps/deepseek-harness/gpt-realtime-2)
- [GitHub Copilot CLI](https://opper.ai/apps/github-copilot-cli/gpt-realtime-2)
- [Aider](https://opper.ai/apps/aider/gpt-realtime-2)
- [Goose](https://opper.ai/apps/goose/gpt-realtime-2)
- [Droid](https://opper.ai/apps/droid/gpt-realtime-2)
- [OpenCode](https://opper.ai/apps/opencode/gpt-realtime-2)
- [Pi](https://opper.ai/apps/pi/gpt-realtime-2)
- [OpenClaw](https://opper.ai/apps/openclaw/gpt-realtime-2)
- [Hermes](https://opper.ai/apps/hermes/gpt-realtime-2)
- [Copilot for Obsidian](https://opper.ai/apps/obsidian-copilot/gpt-realtime-2)
- [OmniRoute](https://opper.ai/apps/omniroute/gpt-realtime-2)
- [Cursor](https://opper.ai/apps/cursor/gpt-realtime-2)
- [Cline](https://opper.ai/apps/cline/gpt-realtime-2)
- [Continue](https://opper.ai/apps/continue/gpt-realtime-2)
- [Kilo Code](https://opper.ai/apps/kilo-code/gpt-realtime-2)
- [GitHub Copilot](https://opper.ai/apps/github-copilot/gpt-realtime-2)
- [Zed](https://opper.ai/apps/zed/gpt-realtime-2)
- [JetBrains AI Assistant](https://opper.ai/apps/jetbrains-ai-assistant/gpt-realtime-2)
- [Voidleap Code](https://opper.ai/apps/voidleap-code/gpt-realtime-2)
- [n8n](https://opper.ai/apps/n8n/gpt-realtime-2)
- [LibreChat](https://opper.ai/apps/librechat/gpt-realtime-2)
- [Open WebUI](https://opper.ai/apps/open-webui/gpt-realtime-2)
- [See all apps](https://opper.ai/apps)

## Other models from OpenAI

- [GPT-6.2 Astra](https://opper.ai/openai/gpt-6-2-astra)
- [GPT-6 Astra](https://opper.ai/openai/gpt-6-astra)
- [GPT-6.2 Sol](https://opper.ai/openai/gpt-6-2-sol)
- [GPT-6.1 Sol](https://opper.ai/openai/gpt-6-1-sol)
- [GPT-6.1 Terra](https://opper.ai/openai/gpt-6-1-terra)
- [GPT-6 Luna](https://opper.ai/openai/gpt-6-luna)
