Tencent Hy4 Launch - Financial Professional Blind Evaluation: leaderboard
Tencent Hy and Tencent Financial Technology blind evaluation: 579 real-world financial tasks; 213 financial professionals, including 153 from external institutions, about 70% with more than five years of experience. Exact grading scale, model efforts, inference settings and backend snapshots are undisclosed. No human performance baseline is reported. Preserve source-run identity and leave theoretical bounds unknown.
Metric: Published expert performance score. Source: www.tencent.com. Saturation forecast: Rough model projection: around 2026. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Hy4 (Benchmark Sweep Run: source=TencentHy4FinanceLaunch; 579 tasks; 213 financial experts; blind evaluation; configuration undisclosed) | 3.04 | |
| 2 | DeepSeek-V4.1-Flash-1017 (Benchmark Sweep Run: source=TencentHy4FinanceLaunch; 579 tasks; 213 financial experts; blind evaluation; configuration undisclosed) | 2.97 | |
| 3 | Kimi K3.1 Pro (Benchmark Sweep Run: source=TencentHy4FinanceLaunch; 579 tasks; 213 financial experts; blind evaluation; configuration undisclosed) | 2.68 | |
| 4 | GLM-5.3 (Benchmark Sweep Run: source=TencentHy4FinanceLaunch; 579 tasks; 213 financial experts; blind evaluation; configuration undisclosed) | 2.68 |
Interactive version: theaggregate.ai/benchmark?slug=tencent-hy4-launch-financial-professional-blind-evaluation · How It Works · Data refreshed daily, snapshot 2026-11-06.