Tau3-Voice English Retail (GPT-5.5 User Simulator): leaderboard

Sierra's τ³-Voice leaderboard (taubench.com), English edition: voice agents (speech-to-speech models, realtime APIs and cascaded speech-to-text/LLM/text-to-speech pipelines) handle τ³-bench customer-service tasks over simulated phone calls, using tools against a domain database, while a simulated customer speaks to them. Custom submissions whose simulated customer runs on GPT-5.5 at xhigh effort instead of the default v1.0 user simulator; the site marks custom rows as not comparable to standard ones, so they are kept on their own boards. Retail domain. Score: pass^1 (%), the share of tasks whose final database state matches the goal. Higher is better.

Metric: Pass^1 (%). Source: taubench.com. Saturation forecast: Rough model projection: around 2026. 6 models tracked.

Top models

Interactive version: theaggregate.ai/benchmark?slug=tau3-voice-english-retail-gpt-5-5-user-simulator · How It Works · Data refreshed daily, snapshot 2026-11-05.