Visual Reasoning on Bongard-OpenWorld
Metric: 2-Class Accuracy (higher is better)
Metric: 2-Class Accuracy (higher is better)
| # | Model↕ | 2-Class Accuracy▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | Gemini-2.0 + CA | 93.6 | No | A Cognitive Paradigm Approach to Probe the Perce... | 2025-01-23 | - |
| 2 | GPT-4o + CA | 92.8 | No | A Cognitive Paradigm Approach to Probe the Perce... | 2025-01-23 | - |
| 3 | Human | 91 | No | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 4 | SNAIL | 64 | Yes | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 5 | InstructBLIP + GPT-4 | 63.8 | No | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 6 | BLIP-2 + ChatGPT (Fine-tuned) | 63.3 | Yes | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 7 | InstructBLIP + ChatGPT + Neuro-Symbolic | 55.5 | No | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 8 | ChatCaptioner + ChatGPT | 49.3 | No | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |
| 9 | Otter | 49.3 | No | Bongard-OpenWorld: Few-Shot Reasoning for Free-f... | 2023-10-16 | Code |