Qwen3 VL 30B A3B Instruct vs Step3 VL 10B
Wins 1 of 7 areas
Long documents
Wins 3 of 7 areas
Reasoning · Images and charts · Following instructions
Step3 VL 10B wins more areas, narrowly.Qwen3 VL 30B A3B Instruct is better at long documents.
Scores updated · 25 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 video13Step3 VL 10B3 of 5 tests · 1 tie
- ReasoningHard problems that need careful thinking01Step3 VL 10B1 of 3 tests · 2 ties
- Following instructionsDoing exactly what it is asked01Step3 VL 10B1 of 1 test
- Long documentsFinding answers in very long texts10Qwen3 VL 30B A3B Instruct1 of 1 test
- CodingWriting and fixing software00Even0 each · 2 ties
- AgentsCarrying out multi-step tasks on its own00Even0 each · 1 tie
- 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 Step3 VL 10B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+17.1points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+10.4points ahead
- Questions that truly need the image to answerMMStar+5.5points ahead
Where Qwen3 VL 30B A3B Instruct pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+2.3points ahead
- Reads text in imagesOCRBench+1.3points ahead
Every test, side by side
All 25 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningStep3 VL 10B
- Humanity's Last ExamStep3 VL 10B by 4.56.310.8+4.5
- GPQA Diamondtie69.569tie
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationStep3 VL 10B by 10.46.316.7+10.4
- AA-Omniscience · AccuracyQwen3 VL 30B A3B Instruct by 2.315.613.3+2.3
Images and chartsStep3 VL 10B
Long documentsQwen3 VL 30B A3B Instruct
- AA-LCRQwen3 VL 30B A3B Instruct by 27.327.30+27.3
Following instructionsStep3 VL 10B
- IFBenchStep3 VL 10B by 17.133.150.2+17.1
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.
- HMMT 2025Step3 VL 10B by 41.550.692.1+41.5
- AIME 2025Step3 VL 10B by 18.469.387.7+18.4
- MATH-VisionStep3 VL 10B by 10.660.270.8+10.6
- ScreenSpot-Pro (No tools)Qwen3 VL 30B A3B Instruct by 8.960.551.5+8.9
- AA-OmniscienceStep3 VL 10B by 4.4-63.4-59+4.4
- AI2DStep3 VL 10B by 4.48589.3+4.4
- AA Agentic IndexQwen3 VL 30B A3B Instruct by 4.19.55.4+4.1
- τ²-Bench Telecom (AA run)Qwen3 VL 30B A3B Instruct by 2.91916.1+2.9
- Artificial Analysis Coding Indextie14.313.9tie
- AA Intelligencetie7.97.7tie
Questions people ask
Which is better, Qwen3 VL 30B A3B Instruct or Step3 VL 10B?
Step3 VL 10B wins three of the seven areas where both have results: reasoning, images and charts and following instructions. Qwen3 VL 30B A3B Instruct wins long documents. They are level on coding, agents and facts.
Which is better for coding?
Neither. All 2 coding tests both models report are ties.
How do you compare the two?
We use the 25 benchmark tests both models have published scores on. The verdict counts the 15 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.