Skip to content

V*

A vision-centric benchmark for high-level multimodal reasoning and perception quality.

12 models

Higher is better.

  1. 1Kimi K2.6Moonshot AI96.9
  2. 2Qwen3.6 PlusAlibaba96.9
  3. 3Qwen3.5 397BAlibaba95.8
  4. 4Step 3.7 FlashStepFun95.3
  5. 5Qwen3.6-27BAlibaba94.7
  6. 6Qwen3.5-27BAlibaba93.7
  7. 7Qwen3.5-122B-A10BAlibaba93.2
  8. 8Qwen3.5-35B-A3BAlibaba92.7
  9. 9Gemini 3 ProGoogle88.0
  10. 10Nemotron 3.5 Super VL 120B A12BNVIDIA85.6
  11. 11GPT-5.2OpenAI75.9
  12. 12Claude Opus 4.5Anthropic67.0
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Vision-centric reasoning benchmark
Tasks
Frontier multimodal reasoning tasks
Authors
Z.AI (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.