Skip to main content
YZ Index

Winzheng · Grounding Rankings

Long document citation verification and grounding accuracy.

Overall Rankings Code Execution Rankings Grounding Rankings Engineering Judgment Rankings Task Communication Rankings Value Rankings Changes
# Model Grounding Code Execution Overall
🥇 DeepSeek V4 Pro DeepSeek
84
67.9 75.2
🥈 Claude Opus 4.7 claude
81.9
85.6 83.9
🥉 GPT-o3 gpt
79.4
76.3 77.7
4 Gemini 3.1 Pro gemini
78.5
73 75.5
5 Gemini 2.5 Pro gemini
78.2
70.5 74
6 Grok 4 grok
77.8
87.8 83.3
7 Claude Sonnet 4.6 claude
75
77.5 76.4
8 Qwen3 Max qwen
71.2
66.6 68.7
9 Doubao Pro doubao
70.7
88.5 80.5
10 GPT-5.5 gpt
69.4
85.4 78.2
11 GLM-4.6 zhipu
24.2
26.3 25.4

Run #342 · Formula v7 · Judge v6.4 · Benchmark v7

Full evaluation: 128 questions from pool of 100yz.rankings.footer.full_eval_suffix · Latest Evaluation2026-09-28 05:02 SGT · Models11models

Winzheng YZ Index — Winzheng · YZ Index