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
101GPT 5.4none· BullshitBench data docs1.3
102Hy 4none· BullshitBench data docs1.2
103Kimi K2.5none· BullshitBench data docs1.2
104Nemotron 3 Super 120B A12Bxhigh· BullshitBench data docs1.2
105GPT 5.5none· BullshitBench data docs1.2
106GPT-6.2 Solmax· BullshitBench data docs1.2
107Muse Spark 1.2minimal· BullshitBench data docs1.2
108GPT-5.6 Terralow· BullshitBench data docs1.2
109GPT 5.5low· BullshitBench data docs1.2
110GPT-5.6 Lunamax· BullshitBench data docs1.2
111GPT-6.2 Sollow· BullshitBench data docs1.2
112GPT 5.4xhigh· BullshitBench data docs1.2
113Muse Spark 1.2xhigh· BullshitBench data docs1.1
114Claude Sonnet 3.7default, thinking· BullshitBench data docs1.1
115Qwen3 Maxnone· BullshitBench data docs1.1
116GPT-5.2 Codexlow· BullshitBench data docs1.1
117Hy4 Flashhigh· BullshitBench data docs1.1
118Gemini 3 Pro Previewlow· BullshitBench data docs1.1
119Claude Opus 4.1none· BullshitBench data docs1.1
120Claude Opus 4.1high· BullshitBench data docs1.1
121Grok 4.3xhigh· BullshitBench data docs1.1
122Nemotron 3 Super 120B A12Bhigh· BullshitBench data docs1.1
123Claude Haiku 3.5default· BullshitBench data docs1.1
124GPT 5.2none· BullshitBench data docs1.1
125GPT-5.5 Promedium· BullshitBench data docs1.1
126Hy4 Flashnone· BullshitBench data docs1.1
127DeepSeek V4 Flash 0731xhigh· BullshitBench data docs1.1
128Claude 3.5 Sonnetdefault· BullshitBench data docs1.1
129GPT-5.5 Proxhigh· BullshitBench data docs1.1
130Nemotron 3 Ultranone· BullshitBench data docs1.1
131GPT-5.6 Lunalow· BullshitBench data docs1.1
132GPT-5.3 Chatdefault· BullshitBench data docs1.1
133GPT-5.1 Chatdefault· BullshitBench data docs1.1
134Gemini 3.1 Pro Previewlow· BullshitBench data docs1
135GPT-5.4 minixhigh· BullshitBench data docs1
136Claude Sonnet 3.7default· BullshitBench data docs1
137Nemotron 3 Super 120B A12Bnone· BullshitBench data docs1
138GPT 5.2high· BullshitBench data docs1
139GPT-5 Codexdefault· BullshitBench data docs1
140GPT-5.2 Codexxhigh· BullshitBench data docs1
141Claude Opus 4default· BullshitBench data docs0.99
142DeepSeek V4 Pro 0813max· BullshitBench data docs0.985
143GPT-5.4 minihigh· BullshitBench data docs0.983
144GPT-5.2 Codexhigh· BullshitBench data docs0.973
145Kimi K2.5high· BullshitBench data docs0.953
146Qwen3.8 Maxmax· BullshitBench data docs0.947
147Gemini 3 Pro Previewhigh· BullshitBench data docs0.943
148GPT-5.4 mininone· BullshitBench data docs0.943
149Nemotron 3 Ultraxhigh· BullshitBench data docs0.938
150MiMo V2.5 Pronone· BullshitBench data docs0.937

Showing 101–150 of 250 results.

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