Skip to main content
BenchLM
Advertise

Benchmark profile

Critical Physics Tasks (CritPt)

A display-only Artificial Analysis metric for research-level physics reasoning.

Data verified

Benchmark score on CritPt — July 30, 2026

BenchLM mirrors the published score view for CritPt. GPT-5.6 Sol leads the public snapshot at 32.3% , followed by GPT-5.5 Pro (30.6%) and GPT-5.6 Terra (30.0%). BenchLM does not use these results to rank models overall.

1Closed

GPT-5.6 Sol

OpenAI

gpt-5-6-sol

32.3%
Overall 81.39Context 1.05M
2Closed

GPT-5.5 Pro

OpenAI

gpt-5-5-pro

30.6%
Overall 62.79Context 1M
3Closed

GPT-5.6 Terra

OpenAI

gpt-5-6-terra

30.0%
Overall 71.99Context 1.05M
158 modelsReasoningCurrentDisplay onlyUpdated July 30, 2026

Benchmark score table (158 models)

Score
1
GPT-5.6 SolOpenAI · Closed
32.3%
2
GPT-5.5 ProOpenAI · Closed
30.6%
3
GPT-5.6 TerraOpenAI · Closed
30.0%
4
GPT-5.4 ProOpenAI · Closed
30.0%
5
Claude Opus 5Anthropic · Closed
29.1%
6
Claude Fable 5Anthropic · Closed
28.6%
7
GPT-5.5OpenAI · Closed
27.1%
8
Gemini 3 Pro Deep ThinkGoogle · Closed
25.7%
9
Kimi K3Moonshot AI · Closed
23.4%
10
GPT-5.4OpenAI · Closed
23.4%
11
Claude Opus 4.8Anthropic · Closed
20.9%
12
GLM-5.2Z.AI · Open weight
20.9%
13
GPT-5.6 LunaOpenAI · Closed
20.6%
14
Gemini 3.1 ProGoogle · Closed
17.7%
15
Claude Sonnet 5Anthropic · Closed
16.9%
16
GPT-5.3 CodexOpenAI · Closed
16.9%
17
GPT-5.3-Codex-SparkOpenAI · Closed
16.9%
18
Grok 4.5xAI · Closed
15.4%
19
Muse Spark 1.1Meta · Closed
15.1%
20
Qwen3.7 MaxAlibaba · Closed
13.4%
21
Gemini 3.5 FlashGoogle · Closed
13.1%
22
DeepSeek V4 Pro (Max)DeepSeek · Open weight
12.9%
23
Claude Opus 4.6 (Adaptive)Anthropic · Closed
12.6%
24
Claude Opus 4.7 (Adaptive)Anthropic · Closed
12.0%
25
GPT-5.2OpenAI · Closed
11.6%
26
Muse SparkMeta · Closed
11.3%
27
Gemini 3.6 FlashGoogle · Closed
10.6%
28
DeepSeek V4 Pro (High)DeepSeek · Open weight
10.0%
29
GPT-5.4 miniOpenAI · Closed
10.0%
30
Kimi K2.7 CodeMoonshot AI · Open weight
10.0%
31
GPT-5.4 nanoOpenAI · Closed
9.3%
32
Qwen3.7 PlusAlibaba · Closed
9.1%
33
Gemini 3 ProGoogle · Closed
9.1%
34
GPT-5.2-CodexOpenAI · Closed
8.7%
35
Kimi K2.6Moonshot AI · Open weight
8.0%
36
Grok 4.3xAI · Closed
8.0%
37
DeepSeek V4 Flash (Max)DeepSeek · Open weight
7.1%
38
GPT-5 (high)OpenAI · Closed
5.7%
39
GPT-5.1-Codex-MaxOpenAI · Closed
5.7%
40
GPT-5.1-CodexOpenAI · Closed
5.7%
41
InklingThinking Machines Lab · Open weight
5.4%
42
Claude Opus 4.7Anthropic · Closed
5.1%
43
GPT-5.1OpenAI · Closed
4.9%
44
Hy3 PreviewTencent · Open weight
4.9%
45
Hy3Tencent · Open weight
4.9%
46
GLM-5.1Z.AI · Open weight
4.6%
47
Claude Opus 4.5 ThinkingAnthropic · Closed
4.6%
48
MiMo-V2.5-ProXiaomi · Closed
4.0%
49
MiniMax M3MiniMax · Open weight
3.7%
50
Qwen 3.6 Max (preview)Alibaba · Closed
3.7%
51
DeepSeek V4 Flash (High)DeepSeek · Open weight
3.4%
52
Kimi K2.5Moonshot AI · Open weight
3.1%
53
Nemotron 3 UltraNVIDIA · Open weight
3.1%
54
Kimi K2.5 (Reasoning)Moonshot AI · Closed
3.1%
55
Nemotron 3 Super 120B A12BNVIDIA · Open weight
3.1%
56
Qwen3.6 PlusAlibaba · Closed
2.9%
57
Grok 4.1 Fast (Reasoning)xAI · Closed
2.9%
58
Grok 4 Fast (Reasoning)xAI · Closed
2.9%
59
Claude Opus 4.6Anthropic · Closed
2.8%
60
Gemini 2.5 ProGoogle · Closed
2.6%
61
Step 3.7 FlashStepFun · Open weight
2.3%
62
Step 3.5 FlashStepFun · Open weight
2.3%
63
GLM-5Z.AI · Open weight
2.0%
64
Grok 4xAI · Closed
2.0%
65
DeepSeek V3.1 (Reasoning)DeepSeek · Open weight
2.0%
66
Qwen3.5 397BAlibaba · Open weight
1.7%
67
GLM-4.7Z.AI · Open weight
1.7%
68
Qwen3.5 397B (Reasoning)Alibaba · Open weight
1.7%
69
Gemini 3 FlashGoogle · Closed
1.4%
70
Gemma 4 31BGoogle · Open weight
1.4%
71
Gemini 2.5 FlashGoogle · Closed
1.4%
72
DeepSeek-R1DeepSeek · Open weight
1.4%
73
GPT-OSS 20BOpenAI · Open weight
1.4%
74
Qwen3.6-27BAlibaba · Open weight
1.1%
75
Claude 4 SonnetAnthropic · Closed
1.1%
76
Gemini 3.1 Flash-LiteGoogle · Closed
1.1%
77
o3OpenAI · Closed
1.1%
78
MiMo-V2-OmniXiaomi · Closed
1.1%
79
GPT-OSS 120BOpenAI · Open weight
1.1%
80
K-ExaoneLG AI Research · Closed
1.1%
81
MiniMax M2.5MiniMax · Closed
1.1%
82
Claude Sonnet 4.6Anthropic · Closed
0.9%
83
Qwen3.5-27BAlibaba · Open weight
0.9%
84
Qwen3.5-35B-A3BAlibaba · Open weight
0.9%
85
DeepSeek V3.2DeepSeek · Open weight
0.9%
86
Trinity-Large-PreviewArcee AI · Open weight
0.9%
87
Trinity-Large-ThinkingArcee AI · Open weight
0.9%
88
Nemotron 3 Nano 30BNVIDIA · Open weight
0.9%
89
Mercury 2Inception · Closed
0.8%
90
Qwen3.5-122B-A10BAlibaba · Open weight
0.6%
91
MiniMax M2.7MiniMax · Open weight
0.6%
92
GLM-5V-TurboZ.AI · Closed
0.6%
93
Gemma 4 E4BGoogle · Open weight
0.6%
94
Claude Opus 4.5Anthropic · Closed
0.3%
95
Qwen3.6-35B-A3BAlibaba · Open weight
0.3%
96
o1OpenAI · Closed
0.3%
97
Command A+Cohere · Open weight
0.3%
98
GLM-5-TurboZ.AI · Closed
0.3%
99
MiMo-V2-ProXiaomi · Closed
0.3%
100
Mistral Small 4Mistral · Open weight
0.3%
101
Sarvam 30BSarvam · Open weight
0.3%
102
Mistral Small 4 (Reasoning)Mistral · Open weight
0.3%
103
GLM-4.7-FlashZ.AI · Open weight
0.3%
104
Exaone 4.0 32BLG AI Research · Open weight
0.0%
105
Gemini 3.5 Flash-LiteGoogle · Closed
0.0%
106
Mistral Medium 3.5 128BMistral · Open weight
0.0%
107
MiMo-V2-FlashXiaomi · Open weight
0.0%
108
Nemotron 3 Nano Omni 30B A3BNVIDIA · Open weight
0.0%
109
LFM2.5-8B-A1BLiquidAI · Open weight
0.0%
110
Kimi K2Moonshot AI · Closed
0.0%
111
Gemma 4 26B A4BGoogle · Open weight
0.0%
112
GPT-4.1 nanoOpenAI · Closed
0.0%
113
GPT-4.1OpenAI · Closed
0.0%
114
Gemma 4 12BGoogle · Open weight
0.0%
115
GPT-4.1 miniOpenAI · Closed
0.0%
116
GLM-4.6Z.AI · Open weight
0.0%
117
DeepSeek V3DeepSeek · Open weight
0.0%
118
GPT-4oOpenAI · Closed
0.0%
119
Llama 4 ScoutMeta · Open weight
0.0%
120
Llama 4 MaverickMeta · Open weight
0.0%
121
Ling 2.6 FlashInclusionAI · Open weight
0.0%
122
Grok 4.1 FastxAI · Closed
0.0%
123
DeepSeek V3.1DeepSeek · Open weight
0.0%
124
Mistral Large 3Mistral · Closed
0.0%
125
GLM-4.5-AirZ.AI · Closed
0.0%
126
Gemma 3 27BGoogle · Open weight
0.0%
127
GPT-5 (medium)OpenAI · Closed
0.0%
128
Mistral Large 2Mistral · Closed
0.0%
129
Llama 3.1 405BMeta · Open weight
0.0%
130
Phi-4Microsoft · Open weight
0.0%
131
Grok Code Fast 1xAI · Closed
0.0%
132
Nemotron Ultra 253BNVIDIA · Open weight
0.0%
133
Claude 3 HaikuAnthropic · Closed
0.0%
134
Claude 4.1 Opus ThinkingAnthropic · Closed
0.0%
135
Nova ProAmazon · Closed
0.0%
136
LFM2.5-VL-1.6B-ExtractLiquidAI · Open weight
0.0%
137
Qwen3 MaxAlibaba · Closed
0.0%
138
Gemma 4 E2BGoogle · Open weight
0.0%
139
Sarvam 105BSarvam · Open weight
0.0%
140
Mistral Medium 3Mistral · Closed
0.0%
141
Granite-4.0-1BIBM · Open weight
0.0%
142
Granite-4.0-350MIBM · Open weight
0.0%
143
Granite-4.0-H-1BIBM · Open weight
0.0%
144
Granite-4.0-H-350MIBM · Open weight
0.0%
145
Exaone 4.0 1.2BLG AI Research · Open weight
0.0%
146
Solar Pro 2Upstage · Closed
0.0%
147
GPT-5 miniOpenAI · Closed
0.0%
148
Ministral 3 14B (Reasoning)Mistral · Open weight
0.0%
149
Ministral 3 14BMistral · Open weight
0.0%
150
GPT-5 nanoOpenAI · Closed
0.0%
151
MiniMax M1 80kMiniMax · Closed
0.0%
152
LFM2-24B-A2BLiquidAI · Closed
0.0%
153
Ministral 3 8B (Reasoning)Mistral · Open weight
0.0%
154
LFM2.5-1.2B-ThinkingLiquidAI · Closed
0.0%
155
Ministral 3 8BMistral · Open weight
0.0%
156
Ministral 3 3B (Reasoning)Mistral · Open weight
0.0%
157
LFM2.5-1.2B-InstructLiquidAI · Closed
0.0%
158
Ministral 3 3BMistral · Open weight
0.0%

The published CritPt snapshot places GPT-5.6 Sol first at 32.3%. The third row is 2.3 points behind. The broader top-10 range is 8.9 points, so many of the published results sit in a relatively narrow band.

158 models have been evaluated on CritPt. The benchmark falls in the Reasoning category. This category carries a 17% weight in BenchLM.ai's overall scoring system. CritPt is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About CritPt

Year

2026

Tasks

Research-level physics questions

Format

Accuracy

Difficulty

Research-level physics reasoning

BenchLM stores CritPt as a display-only research-physics reasoning row from Artificial Analysis' independently evaluated leaderboard.

CritPt Benchmark Leaderboard

BenchLM freshness & provenance

Version

CritPt 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does CritPt measure?

A display-only Artificial Analysis metric for research-level physics reasoning.

Which model scores highest on CritPt?

GPT-5.6 Sol by OpenAI currently leads with a score of 32.3% on CritPt.

How many models are evaluated on CritPt?

158 AI models have been evaluated on CritPt on BenchLM.

Compare Top Models on CritPt

GPT-5.6 Sol vs GPT-5.5 ProGPT-5.5 Pro vs GPT-5.6 TerraGPT-5.6 Terra vs GPT-5.4 ProGPT-5.4 Pro vs Claude Opus 5
Last updated: July 30, 2026 · BenchLM version CritPt 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.