Skip to content

Benchmark

CritPT

Science benchmarks measure how well models answer science questions. Scores are percentages. Higher is better.

Models measured
156
Scores recorded
268
Sources
3
IQ weight
2.67

Score distribution

0.98%
Mean
0.106%
Median
4%
Std dev
0%
10th percentile
0.314%
90th percentile
0%
Min
0%31.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.67 for chat models, measured over 156 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GPT 5.5high· Epoch AI Benchmarking Hub data25.4%
52GPT-6 Solhigh· Epoch AI Benchmarking Hub data25.4%
53GPT-6.1 Sollow· Epoch AI Benchmarking Hub data24.9%
54Muse Spark 1.3max· Epoch AI Benchmarking Hub data24.9%
55Claude Sonnet 5.5high· Epoch AI Benchmarking Hub data24.6%
56GPT-6 Solmedium· Epoch AI Benchmarking Hub data24.6%
57Kimi K3.1 Pro· Epoch AI Benchmarking Hub data24.6%
58DeepSeek V4.1 Flash 1017max· Epoch AI Benchmarking Hub data24%
59GPT 5.4xhigh· Epoch AI Benchmarking Hub data23.4%
60Kimi K3max· Epoch AI Benchmarking Hub data23.4%
61Kimi K3max effort · temp 1.0 · top_p .95· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.23.4%
62Claude Opus 5low· Epoch AI Benchmarking Hub data23.1%
63Gemini 4 Argonmedium· Epoch AI Benchmarking Hub data23.1%
64GPT-5.6 Solmedium· Epoch AI Benchmarking Hub data22.9%
65GPT-5.6 Terrahigh· Epoch AI Benchmarking Hub data22.9%
66Claude Haiku 5.5xhigh· Epoch AI Benchmarking Hub data22.6%
67GPT-6.1 Terramedium· Epoch AI Benchmarking Hub data21.4%
68GLM 5.2setting not named by the source· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.20.9%
69Claude Opus 4.8max· Epoch AI Benchmarking Hub data20.9%
70GLM 5.2max· Epoch AI Benchmarking Hub data20.9%
71Step 5 Preview· Epoch AI Benchmarking Hub data20.9%
72GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data20.6%
73GPT-5.6 Lunaxhigh· Epoch AI Benchmarking Hub data20.6%
74Qwen3.8 Max· Epoch AI Benchmarking Hub data20%
75GPT-6 Astranone· Epoch AI Benchmarking Hub data19.7%
76Grok 4.6xhigh· Epoch AI Benchmarking Hub data19.7%
77GPT-6 Lunamax· Epoch AI Benchmarking Hub data19.4%
78GLM 5.3max· Epoch AI Benchmarking Hub data19.1%
79Claude Haiku 5.5max· Epoch AI Benchmarking Hub data18.9%
80Claude Haiku 5.5high· Epoch AI Benchmarking Hub data18.6%
81DeepSeek V4 Pro 0813high· Epoch AI Benchmarking Hub data18.6%
82GPT 5.5medium· Epoch AI Benchmarking Hub data18.6%
83Gemini 3.8 Flashhigh· Epoch AI Benchmarking Hub data18.3%
84DeepSeek V4 Pro 0813max· Epoch AI Benchmarking Hub data18%
85Grok 4.7high· Epoch AI Benchmarking Hub data18%
86Claude Opus 5.5low· Epoch AI Benchmarking Hub data17.7%
87Gemini 3.1 Pro Preview· Epoch AI Benchmarking Hub data17.7%
88Grok 4.6medium· Epoch AI Benchmarking Hub data17.7%
89Grok 4.7xhigh· Epoch AI Benchmarking Hub data17.7%
90Muse Spark 1.2xhigh· Epoch AI Benchmarking Hub data17.7%
91GPT 5.5Fable/Mythos 5 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.17.7%
92GPT-5.6 Terramedium· Epoch AI Benchmarking Hub data17.4%
93GPT-6 Lunaxhigh· Epoch AI Benchmarking Hub data17.4%
94Grok 4.6high· Epoch AI Benchmarking Hub data17.1%
95Hy4 preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim16.9%
96Hy4 previewofficial· Reported by Tencent Hy· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.16.9%
97Claude Sonnet 5max· Epoch AI Benchmarking Hub data16.9%
98Claude Sonnet 5.5medium· Epoch AI Benchmarking Hub data16.9%
99GPT-5.3 Codexxhigh· Epoch AI Benchmarking Hub data16.9%
100GLM 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim16.7%

Showing 51–100 of 268 results.

Data credits: Data by LLM Stats · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.