Skip to content

Benchmark

BullshitBench v2

Higher scores are better.

Models measured
137
Scores recorded
250
Sources
1
IQ weight
1.14

Score distribution

1
Mean
1.1
Median
0.489
Std dev
0.345
10th percentile
1.6
90th percentile
0.147
Min
0.1471.9
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.14 for chat models, measured over 137 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
151Qwen 3.8 27Bmax· BullshitBench data docs0.937
152Gemini 3.7 Flashminimal· BullshitBench data docs0.933
153Qwen3.8 Flashmax· BullshitBench data docs0.927
154DeepSeek V4 Pro 0813low· BullshitBench data docs0.918
155Gemini 3.6 Flashminimal· BullshitBench data docs0.917
156GPT 5.1default· BullshitBench data docs0.877
157Claude Sonnet 4none· BullshitBench data docs0.87
158Gemini 3.1 Pro Previewhigh· BullshitBench data docs0.867
159MiMo V2.5xhigh· BullshitBench data docs0.867
160Claude Sonnet 4high· BullshitBench data docs0.853
161GPT-5.1 Codexdefault· BullshitBench data docs0.843
162Llama 4 Maverickdefault· BullshitBench data docs0.84
163Qwen3.8 Omni Flashmax· BullshitBench data docs0.837
164GPT-5.2 Chatdefault· BullshitBench data docs0.83
165GPT-5.3 Codexlow· BullshitBench data docs0.83
166Gemini 3.7 Flashmax· BullshitBench data docs0.82
167GPT 5default· BullshitBench data docs0.82
168Gemini 3.6 Flashxhigh· BullshitBench data docs0.783
169GLM 5 Turbohigh· BullshitBench data docs0.78
170GLM 5.2xhigh· BullshitBench data docs0.747
171Gemma 4 31B IThigh· BullshitBench data docs0.74
172GLM 5high· BullshitBench data docs0.727
173Nemotron 3 Nano 30B A3Bnone· BullshitBench data docs0.7
174GPT-5.3 Codexhigh· BullshitBench data docs0.69
175GPT-5.3 Codexxhigh· BullshitBench data docs0.677
176OpenAI o3default· BullshitBench data docs0.667
177DeepSeek V4 Flash 0731none· BullshitBench data docs0.662
178Gemma 4 26B A4Bxhigh· BullshitBench data docs0.647
179GPT-5 Chatdefault· BullshitBench data docs0.637
180MiMo V2.5none· BullshitBench data docs0.633
181Qwen3 Coderdefault· BullshitBench data docs0.633
182GLM 5 Turbonone· BullshitBench data docs0.62
183GLM 5.1xhigh· BullshitBench data docs0.607
184Gemini 2.5 Prodefault· BullshitBench data docs0.597
185Step 3.5 Flashxhigh· BullshitBench data docs0.597
186GLM 5none· BullshitBench data docs0.587
187DeepSeek V4 Flashnone· BullshitBench data docs0.58
188Grok 4.1 Fasthigh· BullshitBench data docs0.577
189DeepSeek V4 Proxhigh· BullshitBench data docs0.573
190Llama 4 Scoutdefault· BullshitBench data docs0.57
191Llama 3.1 8B Instructdefault· BullshitBench data docs0.567
192GLM 5.2none· BullshitBench data docs0.56
193Gemini 3.5 Flashxhigh· BullshitBench data docs0.557
194Gemma 4 31B ITnone· BullshitBench data docs0.557
195Hy3 (Deep Infra)none· BullshitBench data docs0.55
196MiMo V2 Flashnone· BullshitBench data docs0.55
197Gemini 2.5 Flashdefault· BullshitBench data docs0.54
198GPT 5.4 Nanohigh· BullshitBench data docs0.523
199Trinity Large Thinkingminimal· BullshitBench data docs0.51
200GLM 5.1none· BullshitBench data docs0.503

Showing 151–200 of 250 results.

Data credits: BullshitBench by Peter Gostev, MIT (licence)