AI BENCHY Category Failures
General Intelligence: Did not follow instructions
General Intelligence
Did not follow instructions
See which AI models are most likely to hit Did not follow instructions on General Intelligence, so you can spot weak points faster. Sort by: Tests Correct ↑.
Failure Reasons
| Rank | Model | Company | Did not follow instructions Count | Category Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #50 | Gemini 3.1 Flash Lite Preview low | 1 | 4.0 | 0/1 | 1.54s | |
| #53 | Gemini 3.1 Flash Lite high | 1 | 5.0 | 0/1 | 45.7s | |
| #54 | GPT-5 Mini medium | OpenAI | 1 | 4.5 | 0/1 | 13.5s |
| #56 | MiMo-V2.5 medium | Xiaomi | 1 | 5.4 | 0/1 | 5.37s |
| #58 | Gemini 3.1 Flash Lite Preview none | 1 | 4.0 | 0/1 | 741ms | |
| #62 | Step 3.5 Flash medium | Stepfun | 1 | 5.5 | 0/1 | 22.4s |
| #63 | GPT-5.3 Chat none | OpenAI | 1 | 4.6 | 0/1 | 1.99s |
| #64 | MiMo-V2-Flash medium | Xiaomi | 1 | 4.0 | 0/1 | 4.20s |
| #65 | Grok 4.20 medium | X AI | 1 | 3.9 | 0/1 | 24.5s |
| #67 | MiniMax M3 medium | Minimax | 1 | 5.1 | 0/1 | 33.3s |
| #70 | GPT-5.4 Nano medium | OpenAI | 1 | 4.5 | 0/1 | 4.15s |
| #73 | Seed-2.0-Mini medium | Bytedance Seed | 1 | 5.1 | 0/1 | 36.7s |
| #75 | Ring-2.6-1T medium | Inclusionai | 1 | 4.1 | 0/1 | 58.3s |
| #76 | Kimi K2.5 medium | Moonshot AI | 1 | 6.5 | 0/1 | 69.7s |
| #77 | Claude Sonnet 4.6 none | Anthropic | 1 | 6.1 | 0/1 | 2.56s |