BullshitBench v2 Leaderboard
Safety · 2026-02-24
Clear Pushback rate on nonsense prompts; higher scores mean the model rejects more bad claims instead of accepting them.
Top models (higher is better)
| Model | Score |
|---|---|
| Opus 4.8 | 95.0 |
| Sonnet 4.6 | 91.0 |
| Opus 4.5 | 90.0 |
| Opus 4.6 | 87.0 |
| Opus 4.7 | 83.0 |
| Mistral Large 4 Preview | 80.0 |
| Sonnet 5 | 80.0 |
| Sonnet 4.5 | 79.0 |
| Qwen3.8 27B | 79.0 |
| Qwen3.5 397B A17B | 78.0 |
| Claude Fable 5.1 | 77.0 |
| Haiku 4.5 | 77.0 |
| Kimi K3 | 74.0 |
| Claude Opus 5 | 73.0 |
| Qwen3.7-Max (source unspecified) | 72.0 |
| Qwen3.6 Plus (2026-04-02) | 72.0 |
| GPT-6 Astra | 69.0 |
| Grok 4.20 | 67.0 |
| GLM-5.3 | 66.0 |
| Gemini 3.5 Flash-Lite | 66.0 |
| Gemini 4 Argon | 66.0 |
| GPT-6.1 Sol | 65.0 |
| Kimi K2.6 | 65.0 |
| MiniMax M3 | 64.0 |
| Claude Opus 5.5 | 63.0 |
| Qwen3 Max Thinking (2026-01-23) | 63.0 |
| MiMo-V2.5-Pro | 62.0 |
| Claude Sonnet 5.5 | 60.0 |
| GPT-6 Luna Pro | 59.0 |
| GPT-6 Sol | 59.0 |
| GPT-6 Luna | 59.0 |
| Claude Haiku 5.5 | 57.0 |
| GPT-6 Sol Pro | 57.0 |
| DeepSeek V4.1 Flash | 57.0 |
| Muse Wonder | 57.0 |
| Claude Fable 5 | 56.0 |
| Grok 4.5 | 55.0 |
| GPT-5.6 Terra | 54.0 |
| Nemotron 3 Super 120B A12B | 54.0 |
| Kimi K2.5 | 52.0 |