Wikibench

ERQA

reasoning

Embodied Reasoning Question Answering benchmark consisting of 400 multiple-choice visual questions across spatial reasoning, trajectory reasoning, action reasoning, state estimation, and multi-view reasoning for evaluating AI capabilities in physical world interactions

Leaderboard

Showing 19 of 19 results

  1. GPT-5
    65.7%
    i
  2. Qwen3.6 Plus
    65.7%
    i
  3. Qwen3.5-35B-A3B
    64.8%
    i
  4. Muse Spark
    64.7%
    i
  5. o3
    64.0%
    i
  6. Qwen3.6-27B
    62.5%
    i
  7. Qwen3.5-122B-A10B
    62.0%
    i
  8. Qwen3.5-27B
    60.5%
    i
  9. Qwen3 VL 235B A22B Thinking
    52.5%
    i
  10. Qwen3 VL 32B Thinking
    52.3%
    i
  11. Qwen3 VL 235B A22B Instruct
    51.3%
    i
  12. Qwen3 VL 32B Instruct
    48.8%
    i
  13. Qwen3 VL 4B Thinking
    47.3%
    i
  14. Qwen3 VL 8B Thinking
    46.8%
    i
  15. Qwen3 VL 8B Instruct
    45.8%
    i
  16. Qwen3 VL 30B A3B Thinking
    45.3%
    i
  17. Qwen3 VL 30B A3B Instruct
    43.0%
    i
  18. Qwen3 VL 4B Instruct
    41.3%
    i
  19. GPT-4o
    35.2%
    i