ArXivMath

Math reasoning benchmark over arXiv-style problems, reported in the Tencent Hy4 preview model card appendix.

Category

STEM & Reasoning

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for ArXivMath

Model Score Date Verified
Hy3
51.70
28.08.2026 Verified
Hy4 preview
53.60%
28.08.2026 Verified
DeepSeek V4 Pro
37.41%
28.08.2026 Verified
Qwen3.8-2.4T-A95B
55.40%
28.08.2026 Verified
Kimi K3
32.73%
28.08.2026 Verified
Claude Opus 5
71.22%
28.08.2026 Verified
GPT-5.6 Sol
100.00%
28.08.2026 Verified