Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Vision-Language Reasoning on CVBench
Loading...
86.16
Accuracy
Qwen3-VL-8B + GRPO
25.7776
41.4538
57.13
72.8062
Jan 30, 2026
Accuracy
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy
Qwen3-VL-8B + GRPO
Params=8B, Training Sc...
2026.01
86.16
Qwen2-VL-7B + GRPO
Params=7B, Training Sc...
2026.01
75.21
Gemini-3.0-Flash
Zero-shot evaluation=true
2026.01
67.2
Gemini-2.5-Pro
Zero-shot evaluation=true
2026.01
62.4
Qwen2-VL-2B + GRPO
Params=2B, Training Sc...
2026.01
60.31
InternVideo2.5-8B
Params=8B, Training Sc...
2026.01
57.3
LLaVA-Video-7B
Params=7B, Training Sc...
2026.01
52.6
GPT-4V
Params=~1.8T, Training...
2026.01
52.4
Qwen2-VL-7B (baseline)
Params=7B, Training Sc...
2026.01
50.7
Qwen3-VL-8B (baseline)
Params=8B, Training Sc...
2026.01
45.8
Qwen2-VL-2B (baseline)
Params=2B, Training Sc...
2026.01
31.38
Video-LLaVA-7B
Params=7B, Training Sc...
2026.01
28.1
Feedback
Search any
task
Search any
task