Qwen3 30B A3B 2507 Thinking vs Qwen3 32B
Wins 3 of 6 areas
Reasoning · Long documents · Following instructions
Wins 1 of 6 areas
Facts
Qwen3 30B A3B 2507 Thinking wins more areas, narrowly.Qwen3 32B is cheaper and better at facts.
Scores updated · 18 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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.
- ReasoningHard problems that need careful thinking20Qwen3 30B A3B 2507 Thinking2 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts10Qwen3 30B A3B 2507 Thinking1 of 1 test
- Following instructionsDoing exactly what it is asked10Qwen3 30B A3B 2507 Thinking1 of 1 test
- FactsGetting facts right instead of making them up02Qwen3 32B2 of 2 tests
- CodingWriting and fixing software11Even1 each
- AgentsCarrying out multi-step tasks on its own00Even0 each · 3 ties
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 32B costs 69% 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 30B A3B 2507 Thinking pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+14.4points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+3.9points ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+2.9points ahead
Where Qwen3 32B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+4.4points ahead
- Code for real scientific research problemsSciCode+3points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+1.1points ahead
Every test, side by side
All 18 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- SciCodeQwen3 32B by 33336+3
- Terminal-Bench HardQwen3 30B A3B 2507 Thinking by 2.35.33+2.3
AgentsEven
- GDPValtie00tie
- Terminal-Bench 4.0tie00tie
- τ-Bench V3 · Bankingtie5.45.4tie
ReasoningQwen3 30B A3B 2507 Thinking
- GPQA DiamondQwen3 30B A3B 2507 Thinking by 3.970.766.8+3.9
- Humanity's Last ExamQwen3 30B A3B 2507 Thinking by 2.910.37.4+2.9
- CritPttie0.30.3tie
FactsQwen3 32B
- AA-Omniscience · Non-hallucinationQwen3 32B by 4.413.517.9+4.4
- AA-Omniscience · AccuracyQwen3 32B by 1.116.317.4+1.1
Long documentsQwen3 30B A3B 2507 Thinking
- AA-LCRQwen3 30B A3B 2507 Thinking by 61.361.30+61.3
Following instructionsQwen3 30B A3B 2507 Thinking
- IFBenchQwen3 30B A3B 2507 Thinking by 14.450.736.3+14.4
Other results6 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.
- BFCLv4Qwen3 32B by 19.850.570.3+19.8
- AIME 2025Qwen3 30B A3B 2507 Thinking by 12.18572.9+12.1
- AA-OmniscienceQwen3 32B by 5.7-56.1-50.4+5.7
- τ²-Bench Telecom (AA run)Qwen3 32B by 1.728.129.8+1.7
- AA IntelligenceQwen3 30B A3B 2507 Thinking by 1.29.88.6+1.2
- MMLU-Protie81.981.8tie
Questions people ask
Which is better, Qwen3 30B A3B 2507 Thinking or Qwen3 32B?
Qwen3 30B A3B 2507 Thinking wins three of the six areas where both have results: reasoning, long documents and following instructions. Qwen3 32B wins facts, and costs 69% less. They are level on coding and agents.
Which is better for coding?
Neither. They win 1 coding test each of the 2 both models report.
Which is cheaper?
Qwen3 30B A3B 2507 Thinking costs $0.20 per million input tokens and $2.40 per million output tokens; Qwen3 32B costs $0.16 and $0.64. That makes Qwen3 32B about 69% cheaper for the same work.
How do you compare the two?
We use the 18 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 6 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.