Hy3 — benchmark results
Tencent's Hunyuan 3 open MoE model (295B total, 21B active), tuned for agentic tool calling and reasoning. Provider: Tencent. Released 2026-07-06. Access: Open.
Unified ELO 1773 ± 18, rank #165 of 1776 rated models, from 69 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AI Chess Leaderboard (Reasoning) | 1532 | Elo | 95 |
| AA Omniscience - Software Engineering (SWE) - Swift | 76 | Accuracy (%) | 93.9 |
| AA GPQA Diamond | 89.7 | Accuracy (%) | 93.4 |
| Artificial Analysis Intelligence Index | 41.23 | Intelligence Index | 91.9 |
| ZeroEval GPQA Diamond | 90.4 | GPQA Diamond Score | 91.6 |
| AA Humanity's Last Exam | 31.6 | Accuracy (%) | 90.9 |
| AA SciCode | 47.57 | Accuracy (%) | 90.1 |
| AA Omniscience - Software Engineering (SWE) - Rust | 72 | Accuracy (%) | 88.5 |
| AA Omniscience - Software Engineering (SWE) - HTML | 64 | Accuracy (%) | 88.4 |
| AA Omniscience - Software Engineering (SWE) - PHP | 58 | Accuracy (%) | 87.5 |
| AA Omniscience - Software Engineering (SWE) - Java | 38 | Accuracy (%) | 86.4 |
| AA Long Context Reasoning | 66.67 | Accuracy (%) | 86.2 |
Interactive version: theaggregate.ai/model?slug=hy3 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.