Ministral 3 3B vs Qwen3 VL 4B (Reasoning)
Wins 0 of 7 areas
—
Wins 6 of 7 areas
Coding · Agents · Reasoning · Images and charts · Long documents · Following instructions
Qwen3 VL 4B (Reasoning) is the stronger all-rounder.
Scores updated · 23 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.
- CodingWriting and fixing software02Qwen3 VL 4B (Reasoning)2 of 2 tests
- Images and chartsUnderstanding pictures, charts and video02Qwen3 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
- FactsGetting facts right instead of making them up11Even1 each
Agents, long documents and following instructions rest on a single test each.
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 Qwen3 VL 4B (Reasoning) pulls ahead
- Harder college exam questions with imagesMMMU-Pro+13.9points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+13.6points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+9.8points ahead
Where Ministral 3 3B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+11.6points ahead
Every test, side by side
All 23 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 1.815.317.1+1.8
- Terminal-Bench HardQwen3 VL 4B (Reasoning) by 1.501.5+1.5
ReasoningQwen3 VL 4B (Reasoning)
- GPQA DiamondQwen3 VL 4B (Reasoning) by 13.635.849.4+13.6
- Humanity's Last Examtie5.44.6tie
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationMinistral 3 3B by 11.619.88.2+11.6
- AA-Omniscience · AccuracyQwen3 VL 4B (Reasoning) by 3.2912.2+3.2
Images and chartsQwen3 VL 4B (Reasoning)
Long documentsQwen3 VL 4B (Reasoning)
- AA-LCRQwen3 VL 4B (Reasoning) by 4.31721.3+4.3
Following instructionsQwen3 VL 4B (Reasoning)
- IFBenchQwen3 VL 4B (Reasoning) by 9.826.836.6+9.8
Other results11 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.
- OCRBench v2_enQwen3 VL 4B (Reasoning) by 20.441.461.8+20.4
- τ²-Bench Telecom (AA run)Ministral 3 3B by 9.424.915.5+9.4
- DocVQA-valQwen3 VL 4B (Reasoning) by 5.189.694.7+5.1
- AIME 2024Ministral 3 3B by 4.677.572.9+4.6
- AA-OmniscienceMinistral 3 3B by 4.4-64-68.4+4.4
- LiveCodeBenchMinistral 3 3B by 3.554.851.3+3.5
- AIME 2025Qwen3 VL 4B (Reasoning) by 2.472.174.5+2.4
- AA IntelligenceQwen3 VL 4B (Reasoning) by 2.24.87+2.2
- Artificial Analysis Coding IndexQwen3 VL 4B (Reasoning) by 1.94.86.7+1.9
- HallusionBenchMinistral 3 3B by 1.165.264.1+1.1
- MM MTBenchtie7.87.7tie
Questions people ask
Which is better, Ministral 3 3B or Qwen3 VL 4B (Reasoning)?
Qwen3 VL 4B (Reasoning) wins six of the seven areas where both have results: coding, agents, reasoning, images and charts, long documents and following instructions. Ministral 3 3B wins none. They are level on facts.
Which is better for coding?
Qwen3 VL 4B (Reasoning). It wins 2 of the 2 coding tests both models report; Ministral 3 3B wins none.
How do you compare the two?
We use the 23 benchmark tests both models have published scores on. The verdict counts the 12 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.