InternVL3_5-4B wins more areas, narrowly.
Scores updated · 28 tests both models report · How we compare
Where each one wins
Tests won in each of the two areas 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.
Following instructions 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 InternVL3_5-4B pulls ahead
Where InternVL3_5-2B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 28 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Images and chartsInternVL3_5-4B
Other results23 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.
- MMMU (val) (Pass@1)InternVL3_5-4B by 8.75260.7+8.7
- InfographicVQA (val)InternVL3_5-4B by 7.669.376.9+7.6
- MM-IFEvalInternVL3_5-4B by 7.447.154.5+7.4
- MMEInternVL3_5-4B by 7.473.681+7.4
- RealWorldQAInternVL3_5-4B by 6.161.667.7+6.1
- Multilingual MMBenchInternVL3_5-4B by 5.770.976.6+5.7
- MMMBInternVL3_5-4B by 5.276.381.5+5.2
- Multilingual MMMBInternVL3_5-4B by 5.276.381.5+5.2
- MMBenchInternVL3_5-4B by 4.976.281.1+4.9
- ChartQA TestInternVL3_5-4B by 4.581.786.2+4.5
- HallusionBenchInternVL3_5-4B by 4.547.652.1+4.5
- OCRBench v2_enInternVL3_5-4B by 3.645.549.1+3.6
- DocVQA-valInternVL3_5-4B by 3.488.491.8+3.4
- SimpleVQAInternVL3_5-4B by 3.230.533.7+3.2
- IFEvalInternVL3_5-4B by 332.435.4+3
- ScreenSpot-v2 WebInternVL3_5-4B by 2.779.982.6+2.7
- VLM Judge ScoreInternVL3_5-4B by 2.587.790.2+2.5
- ScreenSpot-v2 DesktopInternVL3_5-4B by 2.179.982+2.1
- OCRBench v1InternVL3_5-2B by 1.983.982+1.9
- ScreenSpot-v2 MobileInternVL3_5-4B by 1.686.287.8+1.6
- SEED-Bench (image)InternVL3_5-4B by 175.476.4+1
- POPEtie8888.9tie
- TextVQA-valtie76.677.5tie
Questions people ask
Which is better, InternVL3_5-2B or InternVL3_5-4B?
InternVL3_5-4B wins one of the two areas where both have results: images and charts. InternVL3_5-2B wins none. They are level on following instructions.
How do you compare the two?
We use the 28 benchmark tests both models have published scores on. The verdict counts the 5 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 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.