Skip to content

Benchmark

Dtbench CRI-rescaled score

Scores are percentages.

Models measured
170
Scores recorded
239
Sources
1
IQ weight
1.22

Score distribution

66.3%
Mean
74.2%
Median
27.6%
Std dev
20.9%
10th percentile
95.1%
90th percentile
2.8%
Min
2.8%98.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.22 for chat models, measured over 170 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
151Qwen3 235B A22B Instruct 2507· Epoch AI Benchmarking Hub data64%
152Qwen3.6 27B· Epoch AI Benchmarking Hub data63.6%
153o4 Minihigh· Epoch AI Benchmarking Hub data62.7%
154Claude Sonnet 4· Epoch AI Benchmarking Hub data61.8%
155Qwen3.6 Flash· Epoch AI Benchmarking Hub data61.8%
156Gemini 3.1 Flash Litehigh· Epoch AI Benchmarking Hub data61.3%
157Gemini 2.5 Flash· Epoch AI Benchmarking Hub data60.9%
158GPT OSS 120Bhigh· Epoch AI Benchmarking Hub data60.5%
159Qwen 3 235B A22B· Epoch AI Benchmarking Hub data59.5%
160Mistral Medium2604· Epoch AI Benchmarking Hub data59.1%
161Claude Haiku 5.5none· Epoch AI Benchmarking Hub data58.7%
162Gemma 4 26B A4B· Epoch AI Benchmarking Hub data58.2%
163o1high· Epoch AI Benchmarking Hub data57.8%
164Qwen3.6 35B A3B· Epoch AI Benchmarking Hub data56.5%
165Claude Haiku 4.5· Epoch AI Benchmarking Hub data56%
166Claude Sonnet 5low· Epoch AI Benchmarking Hub data56%
167Qwen3.5 9B· Epoch AI Benchmarking Hub data52%
168Mistral Small2603· Epoch AI Benchmarking Hub data51.6%
169Qwen3 30B A3B2507, thinking· Epoch AI Benchmarking Hub data48.9%
170GPT-5.6 Lunanone· Epoch AI Benchmarking Hub data48.8%
171GPT-4.1 mini· Epoch AI Benchmarking Hub data48%
172o3 Minihigh· Epoch AI Benchmarking Hub data48%
173GPT 4.1· Epoch AI Benchmarking Hub data47.1%
174GPT OSS 20bhigh· Epoch AI Benchmarking Hub data46.7%
175Claude 3.5 Sonnet· Epoch AI Benchmarking Hub data46.4%
176Mistral Medium 3.1· Epoch AI Benchmarking Hub data45.8%
177Qwen3 32B· Epoch AI Benchmarking Hub data45.8%
178Qwen3 30B A3B Instruct· Epoch AI Benchmarking Hub data45.3%
179Grok 2· Epoch AI Benchmarking Hub data42%
180Mistral Large 3· Epoch AI Benchmarking Hub data41.8%
181DeepSeek V3· Epoch AI Benchmarking Hub data41.3%
182GPT 4o· Epoch AI Benchmarking Hub data40.9%
183Qwen3 14B· Epoch AI Benchmarking Hub data40%
184Gemini 2.0 Flash· Epoch AI Benchmarking Hub data38.7%
185Qwen2.5 72B Instruct· Epoch AI Benchmarking Hub data38.2%
186Gemini 2.5 Flash-Litethinking· Epoch AI Benchmarking Hub data38%
187DeepSeek Chat· Epoch AI Benchmarking Hub data37.8%
188GPT 4· Epoch AI Benchmarking Hub data37.8%
189GPT-5 nanohigh· Epoch AI Benchmarking Hub data37.8%
190Mistral Medium 3· Epoch AI Benchmarking Hub data37.1%
191Llama 4 Maverick· Epoch AI Benchmarking Hub data36.5%
192Claude 3 Opus· Epoch AI Benchmarking Hub data36%
193GPT-4 Turbo· Epoch AI Benchmarking Hub data36%
194Llama 3.1 405B Instruct· Epoch AI Benchmarking Hub data35.6%
195C4AI Command A 03 2025· Epoch AI Benchmarking Hub data35.6%
196Magistral Small 2509· Epoch AI Benchmarking Hub data35.5%
197Mistral Large2407· Epoch AI Benchmarking Hub data35.4%
198Mistral Large2411· Epoch AI Benchmarking Hub data34.7%
199Qwen3 30B A3B· Epoch AI Benchmarking Hub data33.8%
200Llama 3.1 70B Instruct· Epoch AI Benchmarking Hub data33.3%

Showing 151–200 of 239 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)