Bongard-OpenWorld is a new benchmark for evaluating real-world few-shot reasoning for machine vision. We hope it can help us better understand the limitations of current visual intelligence and facilitate future research on visual agents with stronger few-shot visual reasoning capabilities.
Variants: Bongard-OpenWorld
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Visual Reasoning | Gemini-2.0 + CA | A Cognitive Paradigm Approach to … | 2025-01-23 |
| Visual Reasoning | GPT-4o + CA | A Cognitive Paradigm Approach to … | 2025-01-23 |
| Visual Reasoning | Human | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | SNAIL | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | InstructBLIP + GPT-4 | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | BLIP-2 + ChatGPT (Fine-tuned) | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | InstructBLIP + ChatGPT + Neuro-Symbolic | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | ChatCaptioner + ChatGPT | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
| Visual Reasoning | Otter | Bongard-OpenWorld: Few-Shot Reasoning for Free-form … | 2023-10-16 |
Recent papers with results on this dataset: