跳到主内容
知行录
leaderboard.cn
Toggle Sidebar
搜索
⌘K
分析
前沿模型
历史趋势
本地模型
资料库
模型库
厂商
基准评测
数据源
显示模式
Toggle Sidebar
知行录
leaderboard.cn
正在加载页面内容
正在加载页面内容
首页
视觉
多模态理解、视觉数学、屏幕与视频任务。
更新于 07/16 02:55
模型
56 个 · 16 项基准
评分
ERQA
VideoMME (w sub.)
Arena Vision Rating
WorldVQA
OmniDocBench 1.5
MMMU-Pro
BabyVision
Artificial Analysis Intelligence Index v4.1
GDP.pdf
MMMU Pro (no tools)
MathVision
VideoMMMU
BenchCAD
BenchCAD (with Python)
MMMU Pro (with tools)
CharXiv-RQ (w. Tool)
1
GPT-5.6 Sol
100.0
-
-
-
-
-
-
-
58.9 pts
30.7%
83%
-
-
70.6%
83.4%
84.6%
-
2
Seed2.1 Pro
95.7
72%
89.2%
-
53%
-
-
73.7%
-
-
-
-
-
-
-
-
85.4%
3
Qwen3.7-Plus
91.6
69.8%
88%
1262.3 Elo
61.1%
91.4%
79%
-
-
-
-
90.3%
85.4%
-
-
-
-
4
GPT-5.4
90.4
-
-
1282.7 Elo
-
89%
81.2%
49.7%
-
-
81.2%
92%
-
-
-
82.1%
-
5
Claude Fable 5
90.0
-
-
1318.1 Elo
-
-
-
-
59.9 pts
29.8%
-
-
-
38.4%
65%
-
-
6
Seed2.1 Turbo
89.1
71.3%
89%
-
48.6%
-
-
62.9%
-
-
-
-
-
-
-
-
82.5%
7
GPT-5.5
88.4
64.5%
89.4%
1286.4 Elo
34.6%
87.5%
81.2%
55.9%
54.8 pts
26%
81.2%
-
86.4%
44.4%
55.8%
83.2%
83.2%
8
Gemini 3.1 Pro
88.3
70.8%
88.4%
1281.5 Elo
56.1%
90%
83%
54.4%
46.5 pts
16.7%
80.5%
89.8%
87.9%
-
-
-
83.5%
9
GPT-5.6 Terra
88.0
-
-
-
-
-
-
-
55 pts
24.7%
80.7%
-
-
62.3%
78.2%
82%
-
10
Qwen3.6-Plus
87.9
65.7%
87.8%
-
33.6%
91.2%
78.8%
-
-
-
-
88%
84%
-
-
-
-
11
Gemini 3 Pro
87.0
70.5%
88.4%
1289.4 Elo
-
88.5%
81%
-
-
-
-
86.6%
87.6%
-
-
-
-
12
Muse Spark 1.1
86.8
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
13
Qwen3.5-397B-A17B
86.2
67.5%
87.5%
1249.8 Elo
-
90.8%
-
-
-
-
-
88.6%
84.7%
-
-
-
-
14
Muse Spark
86.1
-
-
1294.9 Elo
-
-
-
-
-
-
-
-
-
-
-
-
-
15
Claude Opus 4.8
84.5
-
-
1286.4 Elo
-
-
-
-
55.7 pts
22.5%
-
-
-
27.3%
51.8%
-
-
16
Kimi-K2.6
83.7
-
-
1265.6 Elo
-
-
79.4%
39.8%
-
-
-
87.4%
-
-
-
-
-
17
Gemini 3.5 Flash
83.2
-
87.2%
1286.3 Elo
-
-
-
-
50.2 pts
-
-
-
-
-
-
-
-
18
GPT-5.6 Luna
81.7
-
-
-
-
-
-
-
51.2 pts
22.7%
78.4%
-
-
63.1%
73.9%
79.5%
-
19
Claude Sonnet 5
80.5
-
-
1264.5 Elo
-
-
-
-
-
-
-
-
-
-
-
-
-
20
Qwen3.6-27B
80.3
62.5%
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
21
Claude Opus 4.7
80.1
52.5%
-
1306.7 Elo
35.9%
89.3%
77%
22.2%
-
-
-
-
83%
-
-
-
82.1%
22
MiniMax M3
79.6
-
85.4%
1242.5 Elo
-
91.6%
78.1%
-
-
-
-
-
84.6%
-
-
-
-
23
Qwen3.6-35B-A3B
79.1
61.8%
86.6%
-
-
89.9%
75.3%
-
-
-
-
-
83.7%
-
-
-
-
24
MiMo-V2.5
78.7
-
87.7%
1240.2 Elo
-
87.2%
77.9%
-
-
-
-
-
-
-
-
-
-
低
高
单一有效来源基准按 0.5 权重计入,Arena.ai 除外;最高模型归一化为 100