Hy3-preview — benchmark results
Tencent Hunyuan Hy3 preview (295B A21B open MoE). Provider: Tencent. Released 2026-04-24. Access: Open.
Unified ELO 1678 ± 27, rank #309 of 1776 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CLBench | 22.8 | Solving Rate (%) | 91.4 |
| CritPt | 4.6 | Accuracy (self-reported) | 89.3 |
| AI Chess Leaderboard (Reasoning) | 956 | Elo | 80.9 |
| LLM2014 Logic 2026-04 | 56.63 | Median Score | 75 |
| PlanningBench | 36.17 | All-pass (%) (self-reported) | 73.3 |
| LLM2014 Logic 2026-05 | 47.01 | Median Score | 71.1 |
| AA-LCR | 54.7 | Score (self-reported) | 70 |
| Chatbot Arena (Text) | 1412 | Elo | 69.7 |
| SpeechMap Compliance | 76.3 | % Requests Completed | 69.5 |
| LLM2014 Logic 2026-06 | 46.12 | Median Score | 67.5 |
| AI Chess Leaderboard (Continuation) | 699 | Elo | 65.4 |
| SnakeBench | 23.9 | TrueSkill Rating | 59.3 |
Interactive version: theaggregate.ai/model?slug=hy3-preview · How the rankings work · Data refreshed daily, snapshot 2026-07-22.