LFM2.5-VL-1.6B vs Qwen3 VL 4B (Reasoning)
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Qwen3 VL 4B (Reasoning) is the stronger all-rounder.
Scores updated · 24 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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.
- Images and chartsUnderstanding pictures, charts and video03Qwen3 VL 4B (Reasoning)3 of 4 tests · 1 tie
- CodingWriting and fixing software02Qwen3 VL 4B (Reasoning)2 of 2 tests
- FactsGetting facts right instead of making them up02Qwen3 VL 4B (Reasoning)2 of 2 tests
- ReasoningHard problems that need careful thinking01Qwen3 VL 4B (Reasoning)1 of 3 tests · 2 ties
- AgentsCarrying out multi-step tasks on its own01Qwen3 VL 4B (Reasoning)1 of 1 test
- Long documentsFinding answers in very long texts01Qwen3 VL 4B (Reasoning)1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3 VL 4B (Reasoning)1 of 1 test
Agents, long documents and following instructions rest on a single test each.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Qwen3 VL 4B (Reasoning) pulls ahead
- Harder college exam questions with imagesMMMU-Pro+25.5points ahead
- Questions that truly need the image to answerMMStar+22.5points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+20.5points ahead
Where LFM2.5-VL-1.6B pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 24 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 VL 4B (Reasoning)
- SciCodeQwen3 VL 4B (Reasoning) by 14.1317.1+14.1
- Terminal-Bench HardQwen3 VL 4B (Reasoning) by 1.501.5+1.5
ReasoningQwen3 VL 4B (Reasoning)
- GPQA DiamondQwen3 VL 4B (Reasoning) by 20.528.949.4+20.5
- Humanity's Last Examtie5.14.6tie
- CritPttie00tie
FactsQwen3 VL 4B (Reasoning)
- AA-Omniscience · AccuracyQwen3 VL 4B (Reasoning) by 6.45.812.2+6.4
- AA-Omniscience · Non-hallucinationQwen3 VL 4B (Reasoning) by 3.94.38.2+3.9
Images and chartsQwen3 VL 4B (Reasoning)
Long documentsQwen3 VL 4B (Reasoning)
- AA-LCRQwen3 VL 4B (Reasoning) by 21.3021.3+21.3
Following instructionsQwen3 VL 4B (Reasoning)
- IFBenchQwen3 VL 4B (Reasoning) by 3.533.136.6+3.5
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.
- MMMU (val) (Pass@1)Qwen3 VL 4B (Reasoning) by 30.240.670.8+30.2
- OCRBench v2_enQwen3 VL 4B (Reasoning) by 20.341.561.8+20.3
- InfoVQA (val)Qwen3 VL 4B (Reasoning) by 16.862.779.5+16.8
- AA-OmniscienceQwen3 VL 4B (Reasoning) by 16-84.4-68.4+16
- AA Agentic IndexQwen3 VL 4B (Reasoning) by 11.62.814.4+11.6
- RealWorldQAQwen3 VL 4B (Reasoning) by 8.464.873.2+8.4
- DocVQA-valQwen3 VL 4B (Reasoning) by 787.794.7+7
- τ²-Bench Telecom (AA run)Qwen3 VL 4B (Reasoning) by 78.515.5+7
- HallusionBenchQwen3 VL 4B (Reasoning) by 460.164.1+4
- AA IntelligenceQwen3 VL 4B (Reasoning) by 2.24.87+2.2
Questions people ask
Which is better, LFM2.5-VL-1.6B or Qwen3 VL 4B (Reasoning)?
Qwen3 VL 4B (Reasoning) wins all seven areas where both have results: coding, agents, reasoning, facts, images and charts, long documents and following instructions. LFM2.5-VL-1.6B wins none.
Which is better for coding?
Qwen3 VL 4B (Reasoning). It wins 2 of the 2 coding tests both models report; LFM2.5-VL-1.6B wins none.
How do you compare the two?
We use the 24 benchmark tests both models have published scores on. The verdict counts the 14 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.