Reasoning & knowledge

CritPt

Complex Research using Integrated Thinking — Physics Test: 71 unpublished research-level physics challenges (decomposed into 190 checkpoint tasks) across 11 subfields, authored by 50+ working physicists. Reported here as Artificial Analysis's independently-run accuracy on the composite challenges; tool access materially changes results, so runs with and without code execution are not comparable.

How models are tested

Scores are published numbers from the sources below. They are only comparable when the evaluation version, tools, agent harness, and sampling setup match. Missing scores are omitted rather than treated as zero.

Category: Reasoning & knowledge
Metric: Percent
Direction: Higher is better
Coverage: 99 / 318Models in this catalog with a published score
Updated: 2026-09-29Catalog snapshot date
Source: artificialanalysis.ai/evaluations/critpt

Scores

1
GPT-5.6 Sol
OpenAIclosed
32.3%
2
GPT-6 Astra
OpenAIclosed
31.7%
3
GPT-5.5 Pro
OpenAIclosed
30.6%
4
GPT-5.6 Terra
OpenAIclosed
30.0%
5
GPT-5.4 Pro
OpenAIclosed
30.0%
6
Claude Fable 5.1
Anthropicclosed
29.7%
7
Claude Opus 5
Anthropicclosed
29.1%
8
Claude Fable 5
Anthropicclosed
28.6%
9
GPT-5.5
OpenAIclosed
27.1%
10
Muse Spark 1.3
Metaclosed
24.9%
11
GPT-5.4
OpenAIclosed
23.4%
12
Kimi K3
Moonshotopen
23.4%
13
Claude Opus 4.8
Anthropicclosed
20.9%
14
GLM-5.2
Zhipu AIopen
20.9%
15
GPT-5.6 Luna
OpenAIclosed
20.6%
16
GLM-5.3
Zhipu AIopen
19.1%
17
Gemini 3.8 Flash
Googleclosed
18.3%
18
Grok 4.7
xAIclosed
18.0%
19
DeepSeek V4 Pro 0813
DeepSeekclosed
18.0%
20
Muse Spark 1.2
Metaclosed
17.7%
21
Qwen3.8 Max
Alibabaopen
17.7%
22
Grok 4.6
xAIclosed
17.1%
23
Claude Sonnet 5
Anthropicclosed
16.9%
24
DeepSeek V4 Flash 0731
DeepSeekopen
16.6%
25
Grok 4.5
xAIclosed
15.4%
26
GLM-5.3 Flash
Zhipu AIopen
15.4%
27
Muse Spark 1.1
Metaclosed
15.1%
28
Gemini 3.7 Flash
Googleclosed
14.3%
29
DeepSeek V4.1 Flash
DeepSeekopen
14.3%
30
Qwen3.7 Max
Alibabaclosed
13.4%
31
Gemini 3.5 Flash
Googleclosed
13.1%
32
DeepSeek V4 Pro
DeepSeekopen
12.9%
33
Claude Opus 4.6
Anthropicclosed
12.6%
34
Claude Opus 4.7
Anthropicclosed
12.0%
35
Qwen3.8 Flash Next
Alibabaopen
11.1%
36
Gemini 3.6 Flash
Googleclosed
10.6%
37
Kimi K2.7 Code
Moonshotopen
10.0%
38
GPT-5.4 Mini
OpenAIclosed
10.0%
39
GPT-5.4 Nano
OpenAIclosed
9.3%
40
Qwen3.7 Plus
Alibabaclosed
9.1%
41
Gemini 3 Flash
Googleclosed
8.6%
42
Inkling-Small
Thinking Machinesopen
8.3%
43
Grok 4.3
xAIclosed
8.0%
44
Kimi K2.6
Moonshotclosed
8.0%
45
DeepSeek V4 Flash
DeepSeekopen
7.1%
46
GPT-5
OpenAIclosed
5.7%
47
Qwen3.8 27B
Alibabaopen
5.4%
48
Solar Pro 4
Upstageclosed
5.4%
49
Inkling
Thinking Machinesopen
5.4%
50
Hy3
Tencentopen
4.9%
51
Claude Opus 4.5
Anthropicclosed
4.6%
52
K2 Horizon 375B A23B
MBZUAI Institute of Foundation Modelsopen
4.6%
53
GLM-5.1
Zhipu AIopen
4.6%
54
Hunyuan 3 Preview
Tencentclosed
4.6%
55
MiMo-V2.5-Pro
Xiaomiopen
4.0%
56
MiMo-V2.5
Xiaomiopen
3.7%
57
MiniMax M3
MiniMaxopen
3.7%
58
Ring-2.6-1T
InclusionAIopen
3.7%
59
Claude Sonnet 4.6
Anthropicclosed
3.1%
60
Kimi K2.5
Moonshotopen
3.1%
61
Nemotron 3 Ultra
NVIDIAopen
3.1%
62
DeepSeek V3.2
DeepSeekopen
2.9%
63
Qwen3.6 Plus
Alibabaclosed
2.9%
64
MiMo-V2-Flash
Xiaomiopen
2.9%
65
Grok 4 Fast
xAIclosed
2.9%
66
Gemini 2.5 Pro
Googleclosed
2.6%
67
Muse Glimmer
Metaopen
2.6%
68
LongCat-2.0
Meituanopen
2.6%
69
Step-3.5 Flash
StepFunclosed
2.5%
70
Grok 4
xAIclosed
2.0%
71
DeepSeek V3.1
DeepSeekopen
2.0%
72
GLM-5
Zhipu AIopen
2.0%
73
Qwen3.5 397B-A17B
Alibabaopen
1.7%
74
GLM-4.7
Zhipu AIopen
1.7%
75
Ling-3.0 Flash
InclusionAIopen
1.7%
76
Gemma 4 31B
Googleopen
1.4%
77
DeepSeek R1-0528
DeepSeekopen
1.4%
78
gpt-oss-20B
OpenAIopen
1.4%
79
o3
OpenAIclosed
1.1%
80
Gemini 2.5 Flash
Googleclosed
1.1%
81
Qwen3.6 27B
Alibabaopen
1.1%
82
MiniMax M2.5
MiniMaxopen
1.1%
83
gpt-oss-120B
OpenAIopen
1.1%
84
Gemini 3.1 Flash-Lite
Googleclosed
1.1%
85
Claude 3.7 Sonnet
Anthropicclosed
0.9%
86
Trinity Large Thinking
Arcee AIopen
0.9%
87
Mercury 2
Inception Labsclosed
0.8%
88
o4-mini
OpenAIclosed
0.6%
89
DeepSeek R1
DeepSeekopen
0.6%
90
MiniMax M2.7
MiniMaxclosed
0.6%
91
Qwen3.6 35B A3B
Alibabaopen
0.3%
92
Qwen3 32B
Alibabaopen
0.3%
93
Mistral Small 4
Mistralopen
0.3%
94
o1
OpenAIclosed
0.3%
95
Granite 4.2 30B
IBMopen
0.3%
96
Granite 4.2 8B
IBMopen
0.3%
97
GLM-4.7 Flash
Zhipu AIclosed
0.3%
98
Ling-2.6 1T
InclusionAIopen
0.3%
99
INTELLECT-3
Prime Intellectopen
0.3%

Back to all benchmarks