AI BENCHY Category Failures
Combined: Wrong answer
Combined
Wrong answer
See which AI models are most likely to hit Wrong answer on Combined, so you can spot weak points faster. Sort by: Response Time (avg) ↓.
Failure Reasons
| Rank | Model | Company | Wrong answer Count | Category Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #23 | MiMo-V2-Pro medium | Xiaomi | 1 | 4.7 | 0/1 | 64.7s |
| #63 | Qwen3.5-35B-A3B none | Qwen | 1 | 3.0 | 0/1 | 47.4s |
| #70 | Qwen3.5-122B-A10B none | Qwen | 1 | 3.0 | 0/1 | 46.0s |
| #87 | Qwen3 Coder Next none | Qwen | 1 | 3.0 | 0/1 | 45.1s |
| #60 | Gemma 4 26B A4B none | 1 | 3.0 | 0/1 | 30.5s | |
| #73 | Mistral Small 4 medium | Mistral | 1 | 3.0 | 0/1 | 25.3s |
| #88 | Nemotron 3 Super none | NVIDIA | 1 | 3.0 | 0/1 | 20.0s |
| #76 | Kimi K2.5 none | Moonshot AI | 1 | 2.8 | 0/1 | 19.2s |
| #72 | Hunter Alpha none | OpenRouter | 1 | 3.0 | 0/1 | 15.2s |
| #22 | Gemini 3.1 Flash Lite Preview low | 1 | 3.0 | 0/1 | 11.9s | |
| #12 | Gemini 3 PRO Preview medium | 1 | 3.0 | 0/1 | 10.4s | |
| #67 | Qwen3.5-27B none | Qwen | 1 | 2.8 | 0/1 | 9.39s |
| #78 | Trinity Large Preview none | Arcee AI | 1 | 3.0 | 0/1 | 8.91s |
| #89 | GPT-4o-mini none | OpenAI | 1 | 3.0 | 0/1 | 7.58s |
| #49 | Qwen3.5 Plus 2026-02-15 none | Qwen | 1 | 3.0 | 0/1 | 6.65s |