Cognitive Ability Evaluation on CogSense-Bench
82.7Fluid IntelligenceHuman
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Human2026.02 | 82.7 | 91.3 | 88.5 | 97.9 | 78.7 | 88.4 | |
| CogSense-8BParameters=8B2026.02 | 63.8 | 91 | 69 | 68 | 50.5 | 73.8 | |
| Qwen3-VL-30BParameters=30B2026.02 | 30.8 | 34 | 37.2 | 56 | 40.9 | 37.4 | |
| GPT-5.22026.02 | 29.4 | 35.9 | 57.5 | 60 | 37.6 | 40.3 | |
| Gemini 2.5 Flash2026.02 | 23.2 | 40.2 | 31 | 40.2 | 45.3 | 36.3 | |
| Claude Sonnet 42026.02 | 22.5 | 31.3 | 26.6 | 58 | 34.4 | 32.6 | |
| Llama-4-Scout-17BParameters=17B2026.02 | 20.3 | 29.9 | 35.4 | 48.7 | 41.9 | 31.8 | |
| Gemma-3-27BParameters=27B2026.02 | 18.5 | 29.4 | 39.8 | 55.3 | 43 | 32.7 | |
| Grok 4 Fast2026.02 | 13 | 45.4 | 41.6 | 21.3 | 37.6 | 31.7 | |
| Ministral-14BParameters=14B2026.02 | 9.42 | 10.6 | 10.6 | 39.3 | 31.2 | 16.5 | |
| Intern-S12026.02 | 6.6 | 21 | 12.4 | 17.5 | 42.6 | 17.4 | |
| GPT-o32026.02 | 4.7 | 51.4 | 20.4 | 38.7 | 43 | 32.3 |