Official provider results
Hy ASR 3.0 preview launch notes
Tencent’s Hunyuan site describes Hy ASR 3.0 preview as a Hy3-based recogniser with context-aware correction and broad dialect coverage; launch coverage cites word error rates near 3%.
Source snapshots refreshed
Five measurement lanesVOICE / S2S
Reasoning, task completion, conversation dynamics, experience, and latency stay separate.
Benchmark profile
- Scope
- Dialect coverage and context-aware recognition claims from Tencent, plus launch-coverage WER figures
- Primary metric
- Word error rate (lower is better)
- Owner
- Tencent Hunyuan
- Available evidence
- Tencent’s site is descriptive; the WER values come from launch coverage and are marked secondary
What it measures
Voice experience
Is the exchange natural, robust, and responsive?
Available results
- Mandarin WER
- 3.34%
- English WER
- 2.62%
- Dialect coverage
- 10 major regions
Secondary (launch coverage)
Open-source evaluation sets; overall about 3%.
Secondary (launch coverage)
Cantonese 3.12%.
Documented by Tencent
Plus 20+ sub-regions and code-switching.
Interpretation limit
The WER figures are reported by launch coverage (BigGo, August 5, 2026) rather than a Tencent page BenchLM could retrieve; treat them as secondary.