AI BENCHY Category Failures
Data parsing and extraction: Wrong answer
Data parsing and extraction
Wrong answer
See which AI models are most likely to hit Wrong answer on Data parsing and extraction, so you can spot weak points faster. Sort by: Tests Correct ↓.
Failure Reasons
| Rank | Model | Company | Wrong answer Count | Category Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #23 | MiMo-V2-Pro medium | Xiaomi | 1 | 7.3 | 1/2 | 17.2s |
| #54 | Mercury 2 medium | Inception | 1 | 7.3 | 1/2 | 1.11s |
| #64 | DeepSeek V3.2 none | DeepSeek | 1 | 6.3 | 1/2 | 9.42s |
| #68 | gpt-oss-120b medium | OpenAI | 1 | 6.4 | 1/2 | 1.98s |
| #74 | GLM 4.7 Flash none | Z.ai | 1 | 7.3 | 1/2 | 4.82s |
| #76 | Kimi K2.5 none | Moonshot AI | 1 | 7.3 | 1/2 | 42.1s |
| #80 | MiniMax M2.7 medium | Minimax | 1 | 6.3 | 1/2 | 21.9s |
| #81 | Elephant medium | Openrouter | 1 | 6.5 | 1/2 | 979ms |
| #85 | Elephant none | Openrouter | 1 | 6.5 | 1/2 | 1.04s |
| #87 | Qwen3 Coder Next none | Qwen | 1 | 6.5 | 1/2 | 1.32s |
| #91 | Mercury 2 none | Inception | 1 | 7.3 | 1/2 | 667ms |
| #92 | Qwen3 Coder Next medium | Qwen | 1 | 6.5 | 1/2 | 81.8s |
| #96 | GPT-5.4 Nano none | OpenAI | 1 | 6.5 | 1/2 | 1.11s |
| #57 | GPT-5 Nano medium | OpenAI | 2 | 3.7 | 0/2 | 21.4s |
| #71 | MiniMax M2.5 medium | Minimax | 2 | 4.6 | 0/2 | 7.48s |