AI BENCHY Failures
Did not follow instructions Failures
See which AI models run into Did not follow instructions most often, so you can spot reliability risks before choosing one. Sort by: Tests Correct ↑.
| Rank | Model | Company | Did not follow instructions Count | Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #57 | GPT-5 Nano medium | OpenAI | 3 | 6.3 | 7/18 | 44.1s |
| #60 | Gemma 4 26B A4B none | 3 | 6.2 | 7/18 | 6.59s | |
| #62 | Gemini 2.5 Flash none | 1 | 6.2 | 7/18 | 903ms | |
| #63 | Qwen3.5-35B-A3B none | Qwen | 2 | 6.1 | 7/18 | 3.82s |
| #65 | MiMo-V2-Pro none | Xiaomi | 2 | 6.0 | 7/18 | 2.39s |
| #66 | GPT-5.4 none | OpenAI | 1 | 5.9 | 7/18 | 1.51s |
| #68 | gpt-oss-120b medium | OpenAI | 4 | 5.8 | 7/18 | 16.1s |
| #69 | Kimi K2.6 none | Moonshot AI | 3 | 5.8 | 7/18 | 2.05s |
| #50 | Hunter Alpha medium | OpenRouter | 2 | 6.7 | 8/18 | 10.3s |
| #54 | Mercury 2 medium | Inception | 4 | 6.5 | 8/18 | 2.21s |
| #55 | MiMo-V2-Omni none | Xiaomi | 2 | 6.5 | 8/18 | 1.99s |
| #58 | GLM 5V Turbo none | Z.ai | 2 | 6.2 | 8/18 | 3.10s |
| #59 | Qwen3.5-Flash none | Qwen | 1 | 6.2 | 8/18 | 3.25s |
| #44 | GPT-5.4 Mini medium | OpenAI | 5 | 7.3 | 9/18 | 15.2s |
| #45 | GPT-5 Mini medium | OpenAI | 4 | 7.0 | 9/18 | 24.0s |