Skip to content

Benchmark

GPQA

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

Models and systems measured
364
Scores recorded
378
Sources
3
IQ weight
0.75

Score distribution

45.3%
Mean
47.5%
Median
34.4%
Std dev
2.9%
10th percentile
89%
90th percentile
0%
Min
0%97%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.75 for chat models, measured over 359 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
201Mistral Small 24B Instruct2501· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim45.3%
202Mistral Largelatest· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim43.9%
203Gemma 4 E2B-it· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim43.4%
204Gemma 3 27B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim42.4%
205Qwen2 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
42.4%
206Nova Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim42%
207Llama 3.1 70B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
41.7%
208Claude Haiku 3.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim41.6%
209Gemma 3 12B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim40.9%
210Claude 3 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim40.4%
211GPT-4o mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim40.2%
212Nova Micro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim40%
213Gemini 1.5 Flash-8B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim38.4%
214Jamba 1.5 Large· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim36.9%
215phi-3.5-moe-instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
36.8%
216Qwen2.5 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
36.4%
217GPT 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim35.7%
218DeepSeek R1 Distill Qwen 1.5B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
33.8%
219Claude 3 Haiku· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim33.3%
220Llama 3.2 11B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim32.8%
221Llama 3.2 3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
32.8%
222Jamba 1.5 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim32.3%
223Gemma 3 4B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim30.8%
224GPT-3.5 Turbo· LLM Stats data30.8%
225Qwen2.5-Omni 7B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim30.8%
226Llama 3.1 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
30.4%
227L3.3 Ms Nevoria 70b· Hugging Face Open LLM Leaderboard data29.4%
228L3.3 Nevoria R1 70b· Hugging Face Open LLM Leaderboard data29.2%
229gemini-1.0-pro· LLM Stats data27.9%
230Qwen2 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
25.3%
231Phi 4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
25.2%
232Mistral Large Instruct 2411· Hugging Face Open LLM Leaderboard data24.9%
233Gemma 3n E2b It· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim24.8%
234Gemma 3n E4B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim23.7%
235Qwen2.5 32B· Hugging Face Open LLM Leaderboard data21.6%
236EVA-Qwen2.5-72B-v0.2· Hugging Face Open LLM Leaderboard data21.1%
237Qwen2.5 72B· Hugging Face Open LLM Leaderboard data20.7%
238Meta Llama 3 70b· Hugging Face Open LLM Leaderboard data19.7%
239Qwen 2 (72B)· Hugging Face Open LLM Leaderboard data19.2%
240Gemma 3 1b it· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim19.2%
241Qwen2 72B Dracarys· Hugging Face Open LLM Leaderboard data18.8%
242Llama 3.1 70B· Hugging Face Open LLM Leaderboard data18.3%
243Qwen2-VL (72B) Instruct· Hugging Face Open LLM Leaderboard data18.3%
244Magnum V2 72B· Hugging Face Open LLM Leaderboard data18.1%
245arcee-blitz· Hugging Face Open LLM Leaderboard data18%
246Qwen2.5 14B· Hugging Face Open LLM Leaderboard data17.6%
247WizardLM 2 8x22B· Hugging Face Open LLM Leaderboard data17.6%
248mixtral-8x22b-v0.1· Hugging Face Open LLM Leaderboard data16.8%
249Gemma 2 27B IT· Hugging Face Open LLM Leaderboard data16.7%
250Mixtral 8x22B Instruct v0.1· Hugging Face Open LLM Leaderboard data16.4%

Showing 201–250 of 374 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Qwen3.5-397B-A17B-BaseQwen3.5 release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.54.6%
Qwen3.7-Plus-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.5%
Qwen3.8-27B-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.45%
Qwen3.8-Flash-Next-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.4%

Data credits: Data by LLM Stats · 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.