North-Micro-Vision-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.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where North-Micro-Vision-Instruct pulls ahead
- Reads text in imagesOCRBench+5.7points ahead
Where Ministral 3 3B 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 chartsNorth-Micro-Vision-Instruct
- OCRBenchNorth-Micro-Vision-Instruct by 5.773.579.2+5.7
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.
- MMMU DEV_VALMinistral 3 3B by 17.950.832.9+17.9
- MMLU-Pro testMinistral 3 3B by 16.847.530.7+16.8
- MMLU testMinistral 3 3B by 15.66650.4+15.6
- OCRBench v2_enMinistral 3 3B by 4.741.436.7+4.7
- MTL MMBench_DEVMinistral 3 3B by 3.867.463.6+3.8
- HallusionBenchMinistral 3 3B by 3.765.261.5+3.7
- GQA TestDev_BalancedNorth-Micro-Vision-Instruct by 354.457.4+3
- DocVQA-valNorth-Micro-Vision-Instruct by 2.589.692.1+2.5
- ChartQA TestNorth-Micro-Vision-Instruct by 1.779.180.8+1.7
- MMMBtie73.472.8tie
- MMBench DEV_EN_V11tie69.268.7tie
- MMBench_DEVtie0.70.6tie
Questions people ask
Which is better, Ministral 3 3B or North-Micro-Vision-Instruct?
North-Micro-Vision-Instruct wins the one area where both have results: images and charts. Ministral 3 3B 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.