LFM2.5-VL-450M-Extract vs LFM2-VL-450M
Wins 2 of 2 areas
Reasoning · Images and charts
Wins 0 of 2 areas
—
LFM2.5-VL-450M-Extract is the stronger all-rounder.
Scores updated · 14 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.
Reasoning 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 LFM2.5-VL-450M-Extract pulls ahead
- Reads text in imagesOCRBench+2.7points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+2.6points ahead
- Questions that truly need the image to answerMMStar+2.1points ahead
Where LFM2-VL-450M pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 14 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningLFM2.5-VL-450M-Extract
- GPQA DiamondLFM2.5-VL-450M-Extract by 2.625.723.1+2.6
Images and chartsLFM2.5-VL-450M-Extract
Other results10 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.
- MMMBLFM2.5-VL-450M-Extract by 13.868.154.3+13.8
- MM-IFEvalLFM2.5-VL-450M-Extract by 11.94533.1+11.9
- IFEvalLFM2.5-VL-450M-Extract by 9.461.251.8+9.4
- MMVet (Pass@1)LFM2.5-VL-450M-Extract by 7.241.133.9+7.2
- RealWorldQALFM2.5-VL-450M-Extract by 6.458.452+6.4
- MMBench (dev en)LFM2.5-VL-450M-Extract by 4.660.956.3+4.6
- POPELFM2.5-VL-450M-Extract by 3.186.983.8+3.1
- MMLU-ProLFM2.5-VL-450M-Extract by 2.119.317.2+2.1
- MMMU (val) (Pass@1)LFM2-VL-450M by 1.732.734.4+1.7
- InfoVQA (val)LFM2-VL-450M by 1.64344.6+1.6
Questions people ask
Which is better, LFM2.5-VL-450M-Extract or LFM2-VL-450M?
LFM2.5-VL-450M-Extract wins all two areas where both have results: reasoning and images and charts. LFM2-VL-450M wins none.
How do you compare the two?
We use the 14 benchmark tests both models have published scores on. The verdict counts the 4 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 10 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.