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
51Qwen3.5 397B A17Bnone· BullshitBench data docs1.5
52Gemini 3.5 Flash Liteminimal· BullshitBench data docs1.5
53Qwen3 Maxhigh· BullshitBench data docs1.5
54Gemini 4 Argonlow· BullshitBench data docs1.5
55Gemini 4 Hydrogenlow· BullshitBench data docs1.5
56Qwen3.8 Omni Flashnone· BullshitBench data docs1.5
57Gemini 4 Carbonhigh· BullshitBench data docs1.5
58Kimi K2.6none· BullshitBench data docs1.5
59Kimi K3.1 Promax· BullshitBench data docs1.5
60Gemini 4 Argonhigh· BullshitBench data docs1.5
61MiniMax M3xhigh· BullshitBench data docs1.5
62Grok 4.6minimal· BullshitBench data docs1.5
63Gemini 4 Carbonlow· BullshitBench data docs1.5
64Claude Haiku 5.5max· BullshitBench data docs1.5
65GPT-6 Lunamax· BullshitBench data docs1.5
66Claude Haiku 5.5low· BullshitBench data docs1.5
67GPT-6 Solmax· BullshitBench data docs1.5
68Claude Sonnet 5.5max· BullshitBench data docs1.4
69MiniMax M3none· BullshitBench data docs1.4
70Muse Emberminimal· BullshitBench data docs1.4
71GPT-6.2 Astramax· BullshitBench data docs1.4
72Grok 4.20 Multi-Agentxhigh· BullshitBench data docs1.4
73Grok 4.20 Multi-Agentlow· BullshitBench data docs1.4
74GPT-6.1 Terramax· BullshitBench data docs1.4
75MiMo V2.5 Proxhigh· BullshitBench data docs1.4
76DeepSeek V4.1 Flashmax· BullshitBench data docs1.4
77Qwen3.6 Plusxhigh· BullshitBench data docs1.4
78Gemini 3.5 Flash Litexhigh· BullshitBench data docs1.4
79DeepSeek V4.1 Flashlow· BullshitBench data docs1.4
80GPT-6 Solnone· BullshitBench data docs1.4
81GPT-6.1 Terranone· BullshitBench data docs1.4
82Muse Embermax· BullshitBench data docs1.4
83GPT-5.6 Terramax· BullshitBench data docs1.4
84GPT-6 Lunanone· BullshitBench data docs1.4
85GPT-6.2 Astralow· BullshitBench data docs1.4
86Grok 4.6max· BullshitBench data docs1.4
87Hy 4high· BullshitBench data docs1.4
88Qwen3.7 Maxxhigh· BullshitBench data docs1.3
89Muse Wonderminimal· BullshitBench data docs1.3
90Muse Wondermax· BullshitBench data docs1.3
91Grok 4.5low· BullshitBench data docs1.3
92Grok 4.3minimal· BullshitBench data docs1.3
93Grok 4.5high· BullshitBench data docs1.3
94GPT-5.6 Sollow· BullshitBench data docs1.3
95GPT-5.6 Solmax· BullshitBench data docs1.3
96Grok 4.20xhigh· BullshitBench data docs1.3
97Kimi K2.6xhigh· BullshitBench data docs1.3
98GPT 5.5xhigh· BullshitBench data docs1.3
99Grok 4.20low· BullshitBench data docs1.3
100Kimi K3.1 Pronone· BullshitBench data docs1.3

Showing 51–100 of 250 results.

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