Qwen 3 VL 8B (Thinking): benchmark results
Alibaba Qwen 3 VL 8B vision-language model evaluated with thinking enabled. Provider: Alibaba. Released 2025-07-01. Access: Open.
Unified ELO 1511 ± 1, rank #800 of 1761 rated models, from 100 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OCR-Robust | 79.67 | OCR1.0 Clean (self-reported) | 94.1 |
| MedLayXPlain | 64.6 | S (self-reported) | 83.9 |
| JMeetEval | 65.5 | Prompt-level strict accuracy (%) | 83.3 |
| CritPt | 30 | Accuracy (self-reported) | 80.8 |
| AA Omniscience - Software Engineering (SWE) - Rust | 58 | Accuracy (%) | 77.6 |
| AA Omniscience - Software Engineering (SWE) - PHP | 32 | Accuracy (%) | 75.9 |
| JMeetEval - Instruction Level | 80.6 | Instruction-level strict accuracy (%) | 75 |
| MathVision | 59.6 | Overall Accuracy (%) | 74.4 |
| AA Omniscience - Software Engineering (SWE) - Go | 24 | Accuracy (%) | 74.1 |
| CFMME | 53.85 | Average (self-reported) | 73.3 |
| LLM Stats (MuirBench) | 76.8 | Score (%) | 72.7 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 24 | Accuracy (%) | 72.5 |
Interactive version: theaggregate.ai/model?slug=qwen-3-vl-8b-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.