AA Speech to Speech - Tau-Voice 2.0 - Airline: leaderboard

Sierra tau-bench v2.0.0 voice: 50 English tasks each in airline, retail and telecom; three trials; GPT5.5 low simulated caller speaking ElevenLabs eleven_v3; common prompt/tools/policy, 2400-second call limit. State/action/output checks; infrastructure errors excluded and rerun. Overall is equal domain mean. GPT-Live's source-declared backend prompt differs. Separate from V1; no cross-edition score replacement. Metric: Pass@1 (%), mean of three trials. Source: https://artificialanalysis.ai/speech-to-speech

Metric: Pass@1 (%), mean of three trials. Source: artificialanalysis.ai. Saturation forecast: Rough model projection: around 2026. 11 models tracked. Updated 17 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-Live 1 (Astra, medium)84
2Phonic 1High83.33
3GPT-Live 1 (Sol, low)83.33
4Grok Voice 3.071.33
5Qwen Audio 3.0 Realtime Plus64.67
6Eleven Realtime v1 Max63.33
7Gemini 3.8 Live (Extended Thinking)High62.67
8GPT Realtime 2.1High60
9Gemini 3.8 Live58
10Gemini 3.1 Flash Live High54.67
11Eleven Realtime v1 Mini50

Interactive version: theaggregate.ai/benchmark?slug=aa-speech-to-speech-tau-voice-2-0-airline · How It Works · Data refreshed daily, snapshot 2026-11-17.