SShogo Bench

VPCT

Visual physics comprehension: predict where the ball lands.

Agent Index weight
0%
shown for context only
Models tested
24
from the benchmark’s leaderboard and reports
Normalised for the index
0.3333 → 0, 1 → 100
Official site

Results

#ModelScoreNormalisedHarnessCost/taskDate
1Gemini 3 Pro Preview Supported
Google
91.0%source——
2GPT-5.2 Supported
OpenAI
84.0%source——
3Gemini 3 Flash Preview Supported
Google
72.6%source——
4GPT-5 Supported
OpenAI
66.0%source——
5GPT-5.1 Supported
OpenAI
58.7%source——
6o4-mini Estimated
OpenAI
57.5%source——
7o3 Supported
OpenAI
52.0%source——
8Gemini 2.5 Pro Supported
Google
48.0%source——
9Gemini 2.5 Flash Estimated
Google
46.2%source——
10GPT-4.5
OpenAI
45.0%source——
11gemini-robotics-er-1.5-preview
Unknown
40.8%source——
12GPT-5 Mini Estimated
OpenAI
40.2%source——
13Claude Opus 4.5 (latest) Supported
Anthropic
40.0%source——
14GPT-4o (2024-08-06) Supported
OpenAI
40.0%source——
15Claude Sonnet 4.5 (latest) Supported
Anthropic
39.8%source——
16Claude 3.7 Sonnet Supported
Anthropic
39.0%source——
17Claude-Opus-4 Supported
Anthropic
38.0%source——
18GPT-5 Nano
OpenAI
37.2%source——
19o1 Estimated
OpenAI
37.0%source——
20Claude Opus 4.1 (latest) Supported
Anthropic
35.0%source——
21Claude Sonnet 4 Supported
Anthropic
34.0%source——
22GPT-4o mini
OpenAI
34.0%source——
23Claude-Sonnet-3.5 Estimated
Anthropic
33.0%source——
24Gemini 2.5 Flash-Lite
Google
30.0%source——