Skip to content

Benchmark

BullshitBench v2 Clear Pushback Rate (all attempts)

Scores are percentages. Higher is better.

Models measured
137
Scores recorded
250
Sources
1
IQ weight
1.07

Score distribution

41.4%
Mean
43%
Median
24.6%
Std dev
9%
10th percentile
74%
90th percentile
2%
Min
2%95%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.07 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
151GPT-5.4 mininone· BullshitBench data docs32%
152Gemini 3.1 Pro Previewhigh· BullshitBench data docs31%
153Gemini 3.7 Flashmax· BullshitBench data docs31%
154GLM 5 Turbohigh· BullshitBench data docs31%
155GLM 5.2xhigh· BullshitBench data docs31%
156GPT-5.4 minixhigh· BullshitBench data docs31%
157Kimi K2.5high· BullshitBench data docs31%
158Claude Sonnet 4high· BullshitBench data docs30%
159Claude Sonnet 4none· BullshitBench data docs29%
160Llama 4 Maverickdefault· BullshitBench data docs29%
161Qwen 3.8 27Bmax· BullshitBench data docs29%
162Gemini 3.6 Flashxhigh· BullshitBench data docs28%
163GLM 5high· BullshitBench data docs28%
164GPT 5.2high· BullshitBench data docs28%
165Nemotron 3 Nano 30B A3Bnone· BullshitBench data docs28%
166GPT-5.2 Chatdefault· BullshitBench data docs27%
167Qwen3.8 Maxmax· BullshitBench data docs27%
168OpenAI o3default· BullshitBench data docs26%
169Gemma 4 31B IThigh· BullshitBench data docs25%
170GPT 5.1default· BullshitBench data docs25%
171Qwen3.8 Flashmax· BullshitBench data docs25%
172GPT-5.3 Codexlow· BullshitBench data docs24%
173MiMo V2.5none· BullshitBench data docs24%
174GLM 5 Turbonone· BullshitBench data docs23%
175GLM 5.1xhigh· BullshitBench data docs22%
176Step 3.5 Flashxhigh· BullshitBench data docs22%
177Gemma 4 26B A4Bxhigh· BullshitBench data docs21%
178GPT 5default· BullshitBench data docs21%
179DeepSeek V4 Flash 0731none· BullshitBench data docs20%
180Gemini 2.5 Prodefault· BullshitBench data docs20%
181Gemini 3.5 Flashxhigh· BullshitBench data docs20%
182Gemma 4 31B ITnone· BullshitBench data docs20%
183GLM 5none· BullshitBench data docs20%
184GPT-5.3 Codexhigh· BullshitBench data docs20%
185Llama 4 Scoutdefault· BullshitBench data docs20%
186Qwen3 Coderdefault· BullshitBench data docs20%
187Gemini 2.5 Flashdefault· BullshitBench data docs19%
188Gemini 3.5 Flashminimal· BullshitBench data docs19%
189GPT-5.3 Codexxhigh· BullshitBench data docs19%
190Grok 4.1 Fasthigh· BullshitBench data docs19%
191Qwen3.8 Omni Flashmax· BullshitBench data docs19%
192DeepSeek V4 Flashnone· BullshitBench data docs18%
193GLM 5.1none· BullshitBench data docs18%
194GPT-5 Chatdefault· BullshitBench data docs18%
195GLM 5.2none· BullshitBench data docs17%
196Trinity Large Thinkingminimal· BullshitBench data docs17%
197Hy3 (Deep Infra)none· BullshitBench data docs16%
198MiMo V2 Flashnone· BullshitBench data docs16%
199Gemini 2.0 Flashdefault· BullshitBench data docs15%
200DeepSeek V4 Flashxhigh· BullshitBench data docs14%

Showing 151–200 of 250 results.

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