Qwen3 32B is the stronger all-rounder.Phi 4 is cheaper.
Scores updated · 21 tests both models report · How we compare
Where each one wins
Tests won in each of the five 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 thinking02Qwen3 32B2 of 3 tests · 1 tie
- CodingWriting and fixing software01Qwen3 32B1 of 2 tests · 1 tie
- Following instructionsDoing exactly what it is asked01Qwen3 32B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each · 1 tie
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
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.
Phi 4 costs 22% 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 Qwen3 32B pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+12.8points ahead
- Code for real scientific research problemsSciCode+10points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+9.3points ahead
Where Phi 4 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 21 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningQwen3 32B
- GPQA DiamondQwen3 32B by 9.357.566.8+9.3
- Humanity's Last ExamQwen3 32B by 3.63.87.4+3.6
- CritPttie00.3tie
FactsEven
- AA-Omniscience · AccuracyQwen3 32B by 3.314.117.4+3.3
- Vectara HHEM hallucination ratelower is betterPhi 4 by 2.23.75.9+2.2
- AA-Omniscience · Non-hallucinationtie18.817.9tie
Following instructionsQwen3 32B
- IFBenchQwen3 32B by 12.823.536.3+12.8
Other results11 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 29.8029.8+29.8
- LiveBenchQwen3 32B by 27.347.674.9+27.3
- vectara_answer_rateQwen3 32B by 19.280.799.9+19.2
- Arena HardQwen3 32B by 18.475.493.8+18.4
- MMLU-ProQwen3 32B by 11.470.481.8+11.4
- MMLU-ProXQwen3 32B by 1049.959.9+10
- AA-OmniscienceQwen3 32B by 5.3-55.7-50.4+5.3
- Artificial Analysis Coding IndexQwen3 32B by 4.111.215.3+4.1
- AA IntelligenceQwen3 32B by 2.75.98.6+2.7
- vectara_factual_consistencyPhi 4 by 2.296.394.1+2.2
- vectara_avg_summary_lengthtie120.9115.8tie
Questions people ask
Which is better, Phi 4 or Qwen3 32B?
Qwen3 32B wins three of the five areas where both have results: coding, reasoning and following instructions. Phi 4 wins none, but costs 22% less. They are level on facts and long documents.
Which is better for coding?
Qwen3 32B. It wins 1 of the 2 coding tests both models report; Phi 4 wins none, and 1 is a tie.
Which is cheaper?
Phi 4 costs $0.13 per million input tokens and $0.50 per million output tokens; Qwen3 32B costs $0.16 and $0.64. That makes Phi 4 about 22% cheaper for the same work.
How do you compare the two?
We use the 21 benchmark tests both models have published scores on. The verdict counts the 10 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.