SShogo Bench

ProofBench

Write rigorous mathematical proofs.

Agent Index weight
0%
shown for context only
Models tested
84
from the benchmark’s leaderboard and reports
Normalised for the index
0 → 0, 1 → 100
Official site

Results

#ModelScoreNormalisedHarnessCost/taskDate
1Claude Fable 5.1 Supported
Anthropic
100.0%source——
2Claude Fable 5.2 Supported
Anthropic
100.0%source——
3Claude Opus 5.5 Supported
Anthropic
100.0%source——
4Claude Sonnet 5.5 Estimated
Anthropic
100.0%source——
5Claude Opus 5 Supported
Anthropic
99.0%source——
6Gemini 4 Argon Supported
Google
99.0%source——
7GPT-6.1 Sol Estimated
OpenAI
99.0%source——
8GPT-6 Astra Supported
OpenAI
99.0%source——
9Claude Fable 5 Supported
Anthropic
95.0%source——
10Kimi K3.1 Pro
Moonshot AI
91.0%source——
11Claude Haiku 5.5 Estimated
Anthropic
89.0%source——
12Kimi K3 Supported
Moonshot AI
87.0%source——
13Muse Wonder Estimated
Meta
84.0%source——
14GPT-5.6 Sol Supported
OpenAI
83.0%source——
15GPT-6 Sol Estimated
OpenAI
83.0%source——
16Claude Sonnet 5 Supported
Anthropic
77.0%source——
17Tencent Hy4 preview
Tencent
75.0%source——
18GPT-5.6 Terra Estimated
OpenAI
74.0%source——
19MiMo-V2.6-Pro
Xiaomi
70.0%source——
20Claude Opus 4.8 Supported
Anthropic
69.0%source——
21GPT-6 Luna
OpenAI
64.0%source——
22MiMo-V2.6-Flash
Xiaomi
63.0%source——
23GPT-5.6 Luna Estimated
OpenAI
60.0%source——
24Gemini 3.7 Flash Estimated
Google
58.0%source——
25Muse Spark 1.3 Estimated
Meta
58.0%source——
26Qwen3.8 Max Preview Estimated
Alibaba (Qwen)
58.0%source——
27DeepSeek V4 Flash 0731
DeepSeek
56.0%source——
28GPT-5.4 Supported
OpenAI
56.0%source——
29Claude Opus 4.7 Supported
Anthropic
54.0%source——
30DeepSeek V4.1 Flash Estimated
DeepSeek
54.0%source——
31Grok 4.6 Supported
xAI
51.0%source——
32Claude Opus 4.6 Supported
Anthropic
50.0%source——
33DeepSeek V4 Pro 0813
DeepSeek
50.0%source——
34GPT-5.5 Supported
OpenAI
50.0%source——
35GLM-5.3 Supported
Z.ai (Zhipu)
49.0%source——
36Gemini 3.8 Flash Supported
Google
48.0%source——
37Claude Sonnet 4.6 Supported
Anthropic
45.0%source——
38Muse Spark 1.2 Estimated
Meta
43.0%source——
39Step 5 Preview
stepfun
42.0%source——
40Muse Spark 1.1 Estimated
Meta
39.0%source——
41Claude Opus 4.5 (latest) Supported
Anthropic
36.0%source——
42Gemini 3.6 Flash Estimated
Google
36.0%source——
43GLM-5.2 Supported
Z.ai (Zhipu)
35.0%source——
44Gemini 3.5 Flash Supported
Google
31.0%source——
45Grok 4.5 Estimated
xAI
31.0%source——
46Gemini 3.1 Pro Preview Supported
Google
26.0%source——
47Grok 4.7 Estimated
xAI
26.0%source——
48Qwen3.7 Max
Alibaba (Qwen)
26.0%source——
49GLM-5.1 Estimated
Z.ai (Zhipu)
22.2%source——
50MiMo-V2.5-Pro
Xiaomi
22.0%source——
51GLM-5.3-Flash Supported
Z.ai (Zhipu)
21.0%source——
52GPT-5.4 mini
OpenAI
21.0%source——
53Gemini 3 Pro Preview Supported
Google
20.0%source——
54Claude Sonnet 4.5 (latest) Supported
Anthropic
19.0%source——
55Beam 1.0
Reflection
18.0%source——
56GPT-5 Supported
OpenAI
18.0%source——
57MiniMax-M3 Estimated
MiniMax
18.0%source——
58Muse Spark
Meta
17.0%source——
59DeepSeek V4 Pro Estimated
DeepSeek
16.0%source——
60Kimi K2.6 Estimated
Moonshot AI
16.0%source——
61MiMo-V2.5
Xiaomi
16.0%source——
62Qwen3.8 27B
Alibaba (Qwen)
16.0%source——
63Gemini 3 Flash Preview Supported
Google
15.0%source——
64GPT-5.2 Supported
OpenAI
15.0%source——
65Grok 4.20 (Non-Reasoning) Estimated
xAI
14.0%source——
66Gemini 3.5 Flash Lite
Google
13.0%source——
67GPT-5 Nano
OpenAI
12.0%source——
68Grok 4.3 Beta
xAI
11.0%source——
69Mistral Large 4
Mistral
10.0%source——
70GPT-5.1 Codex Max Estimated
OpenAI
9.0%source——
71GPT-5 Mini Estimated
OpenAI
9.0%source——
72Mistral Medium 3.5
Mistral
9.0%source——
73DeepSeek V3.2 Supported
DeepSeek
8.0%source——
74GLM-4.7 Estimated
Z.ai (Zhipu)
6.0%source——
75Inkling Small
thinkingmachines
6.0%source——
76GPT-5.4 nano
OpenAI
5.0%source——
77Grok 4.1 Fast (Non-Reasoning)
xAI
4.0%source——
78MiniMax-M2.5 Estimated
MiniMax
4.0%source——
79Mercury 2.5
inception
3.0%source——
80MiniMax-M2.7
MiniMax
3.0%source——
81Nemotron 3 Ultra
NVIDIA
2.0%source——
82Inkling Estimated
thinkingmachines
0.0%source——
83Laguna M.1
poolside
0.0%source——
84Laguna XS.2
Poolside
0.0%source——