OpenAI GPT-Live-1 API Launch - Full Duplex Bench v3 Response quality: leaderboard
OpenAI's September 2026 GPT-Live-1 API launch evaluation on Full Duplex Bench v 3: the spoken answer to a tool-using request with pauses, hesitations and self-corrections is scored against the reference intent. Response quality in percent; higher is better. GPT-Live-1 uses the reported Terra backend at low effort; the backend generation and Realtime comparators' effort are not specified.
Metric: Response quality (%; spoken answer matches reference intent). Source: openai.com. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-Live-1 (Terra, Low) | 90 | |
| 2 | GPT Realtime 2.1 (API launch, effort unspecified) | 88 | |
| 3 | GPT Realtime 2 (API launch, effort unspecified) | 81 |
Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-live-1-api-launch-full-duplex-bench-v3-response-quality · How It Works · Data refreshed daily, snapshot 2026-11-05.