VECTOR WIREAI INTELLIGENCE
⌘K
PKT
Refresh Models —Deals —Regulatory —Sources —

Questions this page answers

Which is better, Qwen2.5 VL 32B Instruct or Qwen3 VL Thinking (8B)?
Across 12 shared benchmarks, Qwen2.5 VL 32B Instruct scores higher on 2 and Qwen3 VL Thinking (8B) on 10. The widest gap is OSWorld-Verified, where Qwen3 VL Thinking (8B) scores 33.9 against 5.9.

Qwen2.5 VL 32B Instruct vs Qwen3 VL Thinking (8B)

Across 12 shared benchmarks, Qwen2.5 VL 32B Instruct scores higher on 2 and Qwen3 VL Thinking (8B) on 10. The widest gap is OSWorld-Verified, where Qwen3 VL Thinking (8B) scores 33.9 against 5.9.

AlibabavsAlibaba12 shared benchmarks2–10 head-to-head
BenchmarkQwen2.5 VL 32B InstructQwen3 VL Thinking (8B)
CC-OCR77.176.3
GPQA Diamond4669.9
LVBench4955.8
mathvision4062.7
MathVista74.781.4
MMLU-Pro68.877.3
MMMU7073.5
MMMU-Pro49.560.4
MMStar69.575.3
OSWorld-Verified5.933.9
screenspot_pro_no_tools39.446.6
Video-MME77.971.8

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.