Wikibench

MAXIFE

general

MAXIFE is a multilingual benchmark evaluating LLMs on instruction following and execution across multiple languages and cultural contexts.

Leaderboard

Showing 10 of 10 results

  1. Qwen3.7 Max
    89.2%
    i
  2. Qwen3.5-397B-A17B
    88.2%
    i
  3. Qwen3.6 Plus
    88.2%
    i
  4. Qwen3.5-27B
    88.0%
    i
  5. Qwen3.5-122B-A10B
    87.9%
    i
  6. Qwen3.5-35B-A3B
    86.6%
    i
  7. Qwen3.5-9B
    83.4%
    i
  8. Qwen3.5-4B
    78.0%
    i
  9. Qwen3.5-2B
    60.6%
    i
  10. Qwen3.5-0.8B
    39.2%
    i