Qwen3 VL 30B A3B Reasoning vs Qwen3 VL 4B Instruct
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents · Following instructions
Wins 0 of 7 areas
—
Qwen3 VL 30B A3B Reasoning is the stronger all-rounder.
Scores updated · 51 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 video41Qwen3 VL 30B A3B Reasoning4 of 6 tests · 1 tie
- CodingWriting and fixing software30Qwen3 VL 30B A3B Reasoning3 of 3 tests
- ReasoningHard problems that need careful thinking20Qwen3 VL 30B A3B Reasoning2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own20Qwen3 VL 30B A3B Reasoning2 of 2 tests
- FactsGetting facts right instead of making them up20Qwen3 VL 30B A3B Reasoning2 of 2 tests
- Long documentsFinding answers in very long texts10Qwen3 VL 30B A3B Reasoning1 of 1 test
- Following instructionsDoing exactly what it is asked10Qwen3 VL 30B A3B Reasoning1 of 1 test
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 30B A3B Reasoning pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+34.9points ahead
- Reasons across sets of long documentsAA-LCR+28.7points ahead
- Recent programming contest problemsLiveCodeBench v6+26.3points ahead
Where Qwen3 VL 4B Instruct pulls ahead
- Reads text in imagesOCRBench+4.2points ahead
Every test, side by side
All 51 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3 VL 30B A3B Reasoning
- LiveCodeBench v6Qwen3 VL 30B A3B Reasoning by 26.364.237.9+26.3
- SciCodeQwen3 VL 30B A3B Reasoning by 15.128.813.7+15.1
- Terminal-Bench HardQwen3 VL 30B A3B Reasoning by 5.35.30+5.3
AgentsQwen3 VL 30B A3B Reasoning
- GDPValQwen3 VL 30B A3B Reasoning by 8.38.30+8.3
- OSWorld-VerifiedQwen3 VL 30B A3B Reasoning by 4.430.626.2+4.4
ReasoningQwen3 VL 30B A3B Reasoning
- GPQA DiamondQwen3 VL 30B A3B Reasoning by 34.97237.1+34.9
- Humanity's Last ExamQwen3 VL 30B A3B Reasoning by 5.38.93.6+5.3
- CritPttie00tie
FactsQwen3 VL 30B A3B Reasoning
- AA-Omniscience · Non-hallucinationQwen3 VL 30B A3B Reasoning by 7.610.22.6+7.6
- AA-Omniscience · AccuracyQwen3 VL 30B A3B Reasoning by 616.910.9+6
Images and chartsQwen3 VL 30B A3B Reasoning
Long documentsQwen3 VL 30B A3B Reasoning
- AA-LCRQwen3 VL 30B A3B Reasoning by 28.742.714+28.7
Following instructionsQwen3 VL 30B A3B Reasoning
- IFBenchQwen3 VL 30B A3B Reasoning by 13.345.131.8+13.3
Other results33 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.
- HMMT 2025Qwen3 VL 30B A3B Reasoning by 36.967.630.7+36.9
- AIME 2025Qwen3 VL 30B A3B Reasoning by 36.583.146.6+36.5
- SimpleQAQwen3 VL 4B Instruct by 24.123.948+24.1
- VideoMMMUQwen3 VL 30B A3B Reasoning by 18.87556.2+18.8
- AA-OmniscienceQwen3 VL 30B A3B Reasoning by 18.4-57.6-76+18.4
- MMLU-ProXQwen3 VL 30B A3B Reasoning by 16.776.159.4+16.7
- SuperGPQAQwen3 VL 30B A3B Reasoning by 16.156.440.3+16.1
- MATH-VisionQwen3 VL 30B A3B Reasoning by 14.165.751.6+14.1
- MMLU-ProQwen3 VL 30B A3B Reasoning by 13.480.567.1+13.4
- IncludeQwen3 VL 30B A3B Reasoning by 13.174.561.4+13.1
- MMLU-ReduxQwen3 VL 30B A3B Reasoning by 9.490.981.5+9.4
- MMMU (val) (Pass@1)Qwen3 VL 30B A3B Reasoning by 8.67667.4+8.6
- Artificial Analysis Coding IndexQwen3 VL 30B A3B Reasoning by 8.513.14.6+8.5
- HallusionBenchQwen3 VL 30B A3B Reasoning by 8.46657.6+8.4
- CharadesSTAQwen3 VL 30B A3B Reasoning by 7.262.755.5+7.2
- AA Agentic IndexQwen3 VL 30B A3B Reasoning by 7.114.97.8+7.1
- RealWorldQAQwen3 VL 30B A3B Reasoning by 6.577.470.9+6.5
- InfoVQA_testQwen3 VL 30B A3B Reasoning by 5.78680.3+5.7
- BFCL v3Qwen3 VL 30B A3B Reasoning by 5.368.663.3+5.3
- ERQAQwen3 VL 30B A3B Reasoning by 445.341.3+4
- AA IntelligenceQwen3 VL 30B A3B Reasoning by 3.89.55.7+3.8
- τ²-Bench Telecom (AA run)Qwen3 VL 4B Instruct by 3.519.923.4+3.5
- MV-BenchQwen3 VL 30B A3B Reasoning by 3.17268.9+3.1
- LVBenchQwen3 VL 30B A3B Reasoning by 359.256.2+3
- AI2DQwen3 VL 30B A3B Reasoning by 2.886.984.1+2.8
- OCRBench v2 (Chinese)Qwen3 VL 30B A3B Reasoning by 2.860.457.6+2.8
- ScreenSpot-Pro (No tools)Qwen3 VL 4B Instruct by 2.257.359.5+2.2
- CC-OCRQwen3 VL 30B A3B Reasoning by 1.677.876.2+1.6
- OCRBench v2_enQwen3 VL 4B Instruct by 1.162.663.7+1.1
- ScreenSpottie94.794tie
- IFEvaltie81.782.3tie
- MM MTBenchtie7.97.5tie
- DocVQA_testtie9595.3tie
Questions people ask
Which is better, Qwen3 VL 30B A3B Reasoning or Qwen3 VL 4B Instruct?
Qwen3 VL 30B A3B Reasoning wins all seven areas where both have results: coding, agents, reasoning, facts, images and charts, long documents and following instructions. Qwen3 VL 4B Instruct wins none.
Which is better for coding?
Qwen3 VL 30B A3B Reasoning. It wins 3 of the 3 coding tests both models report; Qwen3 VL 4B Instruct wins none.
How do you compare the two?
We use the 51 benchmark tests both models have published scores on. The verdict counts the 18 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 33 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.