DeepSeek-R1 vs Qwen3 32B
Wins 5 of 6 areas
Coding · Agents · Reasoning · Long documents · Following instructions
Wins 1 of 6 areas
Facts
DeepSeek-R1 is the stronger all-rounder.Qwen3 32B is cheaper and better at facts.
Scores updated · 27 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.
- CodingWriting and fixing software30DeepSeek-R13 of 3 tests
- ReasoningHard problems that need careful thinking20DeepSeek-R12 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own10DeepSeek-R11 of 3 tests · 2 ties
- Long documentsFinding answers in very long texts10DeepSeek-R11 of 1 test
- Following instructionsDoing exactly what it is asked10DeepSeek-R11 of 1 test
- FactsGetting facts right instead of making them up12Qwen3 32B2 of 3 tests
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 87% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where DeepSeek-R1 pulls ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+13.9points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+13.1points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+4points ahead
Where Qwen3 32B pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+8.2points ahead
Every test, side by side
All 27 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-R1
- Terminal-Bench 2.1DeepSeek-R1 by 13.919.15.2+13.9
- Terminal-Bench HardDeepSeek-R1 by 3.16.13+3.1
- SciCodeDeepSeek-R1 by 2.338.336+2.3
AgentsDeepSeek-R1
- τ-Bench V3 · BankingDeepSeek-R1 by 16.45.4+1
- GDPValtie00tie
- Terminal-Bench 4.0tie00tie
ReasoningDeepSeek-R1
- GPQA DiamondDeepSeek-R1 by 470.866.8+4
- Humanity's Last ExamDeepSeek-R1 by 1.18.57.4+1.1
- CritPttie0.60.3tie
FactsQwen3 32B
- AA-Omniscience · AccuracyDeepSeek-R1 by 13.130.517.4+13.1
- AA-Omniscience · Non-hallucinationQwen3 32B by 8.29.717.9+8.2
- Vectara HHEM hallucination ratelower is betterQwen3 32B by 5.411.35.9+5.4
Following instructionsDeepSeek-R1
- IFBenchDeepSeek-R1 by 2.73936.3+2.7
Other results13 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.
- τ²-Bench Telecom (AA run)Qwen3 32B by 18.411.429.8+18.4
- AA-OmniscienceDeepSeek-R1 by 18.2-32.2-50.4+18.2
- MMLU-ProXDeepSeek-R1 by 15.675.559.9+15.6
- Aider-PolyglotDeepSeek-R1 by 13.353.340+13.3
- Artificial Analysis Coding IndexDeepSeek-R1 by 9.324.615.3+9.3
- vectara_factual_consistencyQwen3 32B by 5.488.794.1+5.4
- AIME 2025Qwen3 32B by 2.97072.9+2.9
- vectara_answer_rateQwen3 32B by 2.99799.9+2.9
- AA IntelligenceDeepSeek-R1 by 2.811.48.6+2.8
- vectara_avg_summary_lengthQwen3 32B by 22.3 rating points93.5115.8+22.3 rating
- LiveCodeBenchQwen3 32B by 2.263.565.7+2.2
- MMLU-ProDeepSeek-R1 by 2.28481.8+2.2
- AIME 2024Qwen3 32B by 1.679.881.4+1.6
Questions people ask
Which is better, DeepSeek-R1 or Qwen3 32B?
DeepSeek-R1 wins five of the six areas where both have results: coding, agents, reasoning, long documents and following instructions. Qwen3 32B wins facts, and costs 87% less.
Which is better for coding?
DeepSeek-R1. It wins 3 of the 3 coding tests both models report; Qwen3 32B wins none.
Which is cheaper?
DeepSeek-R1 costs $2.00 per million input tokens and $4.00 per million output tokens; Qwen3 32B costs $0.16 and $0.64. That makes Qwen3 32B about 87% cheaper for the same work.
How do you compare the two?
We use the 27 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 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.