Qwen 3 VL 32B Instruct — benchmark results

Alibaba's open Qwen 3 VL 32B instruct vision-language model. Provider: Alibaba. Released 2025-07-01. Access: Open.

Unified ELO 1546 ± 21, rank #634 of 1776 rated models, from 102 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
FIND77.8Zero shot en (self-reported)100
LLM Stats (MLVU-M)82.1Score (%)100
LLM Stats (ScreenSpot)95.8Score (%)100
LLM Stats (OCRBench-V2 (en))67.4Score (%)90.9
LLM Stats (DocVQAtest)96.9Score (%)90
LLM Stats (CharXiv-D)90.5Score (%)86.7
CulMind46.6S (self-reported)84.6
VANTAGE-Bench - Video QA71.3Video QA (%)84.6
TriViewBench61.38Overall (self-reported)83.3
MechVQA76.5Total (self-reported)81.2
Physical AI Bench - Understanding Overall62Overall Score (%)79.2
VANTAGE-Bench55.05Overall (%)76.9

Interactive version: theaggregate.ai/model?slug=qwen-3-vl-32b-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.