Qwen3 VL 2B Instruct is the stronger all-rounder.
Scores updated · 13 tests both models report · How we compare
Where each one wins
Tests won in each of the one area where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
Images and charts rests on a single test.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Qwen3 VL 2B Instruct pulls ahead
- Reads text in imagesOCRBench+10.9points ahead
Where Phi 3.5 Vision Instruct pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 13 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Images and chartsQwen3 VL 2B Instruct
- OCRBenchQwen3 VL 2B Instruct by 10.964.275.1+10.9
Other results12 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- MMLU testQwen3 VL 2B Instruct by 27.535.563+27.5
- MMLU-Pro testQwen3 VL 2B Instruct by 14.228.642.8+14.2
- ChartQA TestPhi 3.5 Vision Instruct by 12.882.169.3+12.8
- HallusionBenchQwen3 VL 2B Instruct by 8.858.567.3+8.8
- GQA TestDev_BalancedPhi 3.5 Vision Instruct by 7.86557.2+7.8
- OCRBench v2_enQwen3 VL 2B Instruct by 7.833.941.7+7.8
- MMMU DEV_VALPhi 3.5 Vision Instruct by 5.343.237.9+5.3
- MTL MMBench_DEVQwen3 VL 2B Instruct by 4.561.966.4+4.5
- MMMBQwen3 VL 2B Instruct by 3.768.672.3+3.7
- DocVQA-valPhi 3.5 Vision Instruct by 3.58682.5+3.5
- MMBench DEV_EN_V11Qwen3 VL 2B Instruct by 1.373.174.4+1.3
- MMBench_DEVtie0.60.7tie
Questions people ask
Which is better, Phi 3.5 Vision Instruct or Qwen3 VL 2B Instruct?
Qwen3 VL 2B Instruct wins the one area where both have results: images and charts. Phi 3.5 Vision Instruct wins none.
How do you compare the two?
We use the 13 benchmark tests both models have published scores on. The verdict counts the 1 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.