Qwen3-VL-8B-Instruct — Real Scores Across 3 Benchmarks (2026)

benchmark.space

benchmarks rankings compare voices transcripts papers articles

arxiv

Qwen3-VL-8B-Instruct

3 benchmarks

DocVQA

#14 of 15

89.2%

MathVista

#8 of 13

77.1%

MMMU

#17 of 24

60.7%

benchmark.space canonical public ledger of AI capability