V*

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

In index
Multimodal
Best in multimodal

Models14

Top result96.9%Kimi K2.6

Top-3 spread1.1 pts

Year—

Result and price

11 of 14 results

Results

14 results

#ModelResult

01Kimi K2.696.9%

01Qwen3.6 Plus96.9%

03Qwen3.5 397B95.8%

04Step 3.7 Flash95.3%

05Qwen3.6-27B94.7%

06Qwen3.5-27B93.7%

07Qwen3.5-122B-A10B93.2%

08Qwen3.5-35B-A3B92.7%

09Gemini 3 Pro88.0%

10LFM3-VL-2B86.4%

11Nemotron 3.5 Super VL 120B A12B NVFP486.0%

12Nemotron 3.5 Super VL 120B A12B85.6%

13GPT-5.275.9%

14Claude Opus 4.567.0%