Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d271Policy30d66, 14 pendingSources1,015

Qwen3 VL 30B A3B Instruct vs Qwen3 VL 30B A3B Reasoning

Alibaba · released

Wins 0 of 7 areas

—

Alibaba · released

Wins 5 of 7 areas

Agents · Reasoning · Facts · Long documents · Following instructions

Qwen3 VL 30B A3B Reasoning is the stronger all-rounder.Qwen3 VL 30B A3B Instruct is cheaper.

Scores updated · 53 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.

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.

Qwen3 VL 30B A3B Instruct$0.20 to read · $0.80 to write$1.00Price from Artificial Analysis
Qwen3 VL 30B A3B Reasoning$0.20 to read · $2.40 to write$2.60Price from Alibaba's own price page

Qwen3 VL 30B A3B Instruct costs 62% less for the same work.

The biggest differences

The three tests each model wins by the widest margin. Scores are out of 100.

Where Qwen3 VL 30B A3B Reasoning pulls ahead

  • Recent programming contest problemsLiveCodeBench v6+21.6points aheadQwen3 VL 30B A3B Reasoning64.2Qwen3 VL 30B A3B Instruct42.6
  • Reasons across sets of long documentsAA-LCR+15.4points aheadQwen3 VL 30B A3B Reasoning42.7Qwen3 VL 30B A3B Instruct27.3
  • Follows unfamiliar, precisely checkable instructionsIFBench+12points aheadQwen3 VL 30B A3B Reasoning45.1Qwen3 VL 30B A3B Instruct33.1

Where Qwen3 VL 30B A3B Instruct pulls ahead

  • Reads text in imagesOCRBench+6.4points aheadQwen3 VL 30B A3B Instruct90.3Qwen3 VL 30B A3B Reasoning83.9
  • Visual perception tasks people solve at a glanceBLINK+2.3points aheadQwen3 VL 30B A3B Instruct67.7Qwen3 VL 30B A3B Reasoning65.4
  • Code for real scientific research problemsSciCode+2points aheadQwen3 VL 30B A3B Instruct30.8Qwen3 VL 30B A3B Reasoning28.8

Every test, side by side

All 53 tests both models report. The winning score is in its model's colour; marks a score checked independently.

CodingEven

Full coding ranking

AgentsQwen3 VL 30B A3B Reasoning

Full agents ranking

ReasoningQwen3 VL 30B A3B Reasoning

Full reasoning ranking

FactsQwen3 VL 30B A3B Reasoning

Full facts ranking

Images and chartsEven
  • CharXiv (RQ)Qwen3 VL 30B A3B Reasoning by 7.748.956.6+7.7
  • OCRBenchQwen3 VL 30B A3B Instruct by 6.490.383.9+6.4
  • MMStarQwen3 VL 30B A3B Reasoning by 3.472.175.5+3.4
  • BLINKQwen3 VL 30B A3B Instruct by 2.367.765.4+2.3
  • MathVistaQwen3 VL 30B A3B Reasoning by 1.880.181.9+1.8
  • Video-MMEQwen3 VL 30B A3B Instruct by 1.274.573.3+1.2
  • MMMU-Protie62.161.8tie

Full images and charts ranking

Long documentsQwen3 VL 30B A3B Reasoning
  • AA-LCRQwen3 VL 30B A3B Reasoning by 15.427.342.7+15.4

Full long documents ranking

Following instructionsQwen3 VL 30B A3B Reasoning
  • IFBenchQwen3 VL 30B A3B Reasoning by 1233.145.1+12

Full following instructions ranking

Other results34 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.

Questions people ask

Which is better, Qwen3 VL 30B A3B Instruct or Qwen3 VL 30B A3B Reasoning?

Qwen3 VL 30B A3B Reasoning wins five of the seven areas where both have results: agents, reasoning, facts, long documents and following instructions. Qwen3 VL 30B A3B Instruct wins none, but costs 62% less. They are level on coding and images and charts.

Which is better for coding?

Neither. They win 1 coding test each of the 3 both models report, and 1 is a tie.

Which is cheaper?

Qwen3 VL 30B A3B Instruct costs $0.20 per million input tokens and $0.80 per million output tokens; Qwen3 VL 30B A3B Reasoning costs $0.20 and $2.40. That makes Qwen3 VL 30B A3B Instruct about 62% cheaper for the same work.

How do you compare the two?

We use the 53 benchmark tests both models have published scores on. The verdict counts the 19 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 34 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.