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

BenchmarkScoreMetricPercentile
AI Chess Leaderboard (Reasoning)1532Elo95
AA Omniscience - Software Engineering (SWE) - Swift76Accuracy (%)93.9
AA GPQA Diamond89.7Accuracy (%)93.4
Artificial Analysis Intelligence Index41.23Intelligence Index91.9
ZeroEval GPQA Diamond90.4GPQA Diamond Score91.6
AA Humanity's Last Exam31.6Accuracy (%)90.9
AA SciCode47.57Accuracy (%)90.1
AA Omniscience - Software Engineering (SWE) - Rust72Accuracy (%)88.5
AA Omniscience - Software Engineering (SWE) - HTML64Accuracy (%)88.4
AA Omniscience - Software Engineering (SWE) - PHP58Accuracy (%)87.5
AA Omniscience - Software Engineering (SWE) - Java38Accuracy (%)86.4
AA Long Context Reasoning66.67Accuracy (%)86.2

Interactive version: theaggregate.ai/model?slug=hy3 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.