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
151Inkling FP4xhigh· Epoch AI Benchmarking Hub data5.4%
152Qwen 3.8 27Bxhigh· Epoch AI Benchmarking Hub data5.4%
153GPT-5.6 Solnone· Epoch AI Benchmarking Hub data5.1%
154GPT 5.1· Epoch AI Benchmarking Hub data4.9%
155GPT-5.6 Lunamedium· Epoch AI Benchmarking Hub data4.9%
156GPT-6.1 Terranone· Epoch AI Benchmarking Hub data4.9%
157Claude Sonnet 5low· Epoch AI Benchmarking Hub data4.6%
158GLM 5.1· Epoch AI Benchmarking Hub data4.6%
159Gemini 3.8 Flashlow· Epoch AI Benchmarking Hub data4%
160GPT-6 Solnone· Epoch AI Benchmarking Hub data4%
161MiMo V2.5 Pro· Epoch AI Benchmarking Hub data4%
162MiMo V2.5· Epoch AI Benchmarking Hub data3.7%
163MiniMax M3· Epoch AI Benchmarking Hub data3.7%
164Ring 2.6 1T· Epoch AI Benchmarking Hub data3.7%
165DeepSeek V4 Flashhigh· Epoch AI Benchmarking Hub data3.4%
166Claude Sonnet 4.6max· Epoch AI Benchmarking Hub data3.1%
167GLM 5.2none· Epoch AI Benchmarking Hub data3.1%
168Kimi K2.5· Epoch AI Benchmarking Hub data3.1%
169Kimi K3low· Epoch AI Benchmarking Hub data3.1%
170Nemotron 3 Super· Epoch AI Benchmarking Hub data3.1%
171Nemotron 3 Ultra· Epoch AI Benchmarking Hub data
Values from 2 sources
3.1%
172Nemotron 3 UltraNeMo Evaluator SDK · no tools· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.3.1%
173DeepSeek V3.2 Expthinking· Epoch AI Benchmarking Hub data2.9%
174Qwen3.6 Plus· Epoch AI Benchmarking Hub data2.9%
175GPT-5.6 Lunalow· Epoch AI Benchmarking Hub data2.6%
176GPT-6 Lunalow· Epoch AI Benchmarking Hub data2.6%
177Muse Glimmerhigh· Epoch AI Benchmarking Hub data2.6%
178Step 3.7 Flash· Epoch AI Benchmarking Hub data2.3%
179Gemini 2.5 Pro· Epoch AI Benchmarking Hub data2%
180GPT-5.6 Terranone· Epoch AI Benchmarking Hub data2%
181DeepSeek V3.1 Terminus· Epoch AI Benchmarking Hub data1.7%
182GLM 4.7· Epoch AI Benchmarking Hub data1.7%
183Gemma 4 31B IT· Epoch AI Benchmarking Hub data1.4%
184GPT 5.5none· Epoch AI Benchmarking Hub data1.4%
185GPT OSS 20bhigh· Epoch AI Benchmarking Hub data1.4%
186Claude Sonnet 4.5· Epoch AI Benchmarking Hub data1.1%
187Claude Sonnet 5high· Epoch AI Benchmarking Hub data1.1%
188Gemini 3.1 Flash Lite· Epoch AI Benchmarking Hub data1.1%
189GLM 4.6· Epoch AI Benchmarking Hub data1.1%
190GPT OSS 120Bhigh· Epoch AI Benchmarking Hub data1.1%
191GPT-6 Lunanone· Epoch AI Benchmarking Hub data1.1%
192OpenAI o3high· Epoch AI Benchmarking Hub data1.1%
193DeepSeek R1· Epoch AI Benchmarking Hub data1.1%
194Gemini 2.5 Flash· Epoch AI Benchmarking Hub data1.1%
195Qwen3.5 122B-A10Bnone· Epoch AI Benchmarking Hub data0.857%
196Qwen3.5 397B A17Bnone· Epoch AI Benchmarking Hub data0.857%
197Qwen3.6 27Bnone· Epoch AI Benchmarking Hub data0.857%
198Trinity Large Thinking· Epoch AI Benchmarking Hub data0.857%
199Mercury 2· Epoch AI Benchmarking Hub data0.849%
200o4 Minihigh· Epoch AI Benchmarking Hub data0.6%

Showing 151–200 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.