GPT-4o mini vs Qwen2.5 Instruct 72B
Wins 0 of 3 areas
—
Wins 3 of 3 areas
Coding · Reasoning · Following instructions
Qwen2.5 Instruct 72B is the stronger all-rounder.GPT-4o mini is cheaper.
Scores updated · 25 tests both models report · How we compare
Where each one wins
Tests won in each of the three 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.
Following instructions rests on a single test.
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.
GPT-4o mini costs 23% 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 Qwen2.5 Instruct 72B pulls ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+15.1points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+6.5points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+5.9points ahead
Where GPT-4o mini pulls ahead
No clear win on a test scored out of 100.
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.
CodingQwen2.5 Instruct 72B
- SWE-bench VerifiedQwen2.5 Instruct 72B by 15.18.723.8+15.1
- SciCodeQwen2.5 Instruct 72B by 3.822.926.7+3.8
ReasoningQwen2.5 Instruct 72B
- GPQA DiamondQwen2.5 Instruct 72B by 6.542.649.1+6.5
- Humanity's Last Examtie4.23.6tie
Following instructionsQwen2.5 Instruct 72B
- IFBenchQwen2.5 Instruct 72B by 5.93136.9+5.9
Other results20 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.
- LiveCodeBenchQwen2.5 Instruct 72B by 2035.555.5+20
- HarmBenchGPT-4o mini by 12.184.972.8+12.1
- GSM8KQwen2.5 Instruct 72B by 11.584.395.8+11.5
- MGSMGPT-4o mini by 10.88776.2+10.8
- MMLUQwen2.5 Instruct 72B by 10.266.877+10.2
- MMLU-ProQwen2.5 Instruct 72B by 9.461.771.1+9.4
- MATHQwen2.5 Instruct 72B by 8.280.288.4+8.2
- OpenBookQAQwen2.5 Instruct 72B by 4.29296.2+4.2
- MBPPQwen2.5 Instruct 72B by 3.484.888.2+3.4
- DROPGPT-4o mini by 379.776.7+3
- AIR-Bench 2024Qwen2.5 Instruct 72B by 2.756.359+2.7
- naturalquestions_closedbookGPT-4o mini by 2.738.535.9+2.7
- NarrativeQAGPT-4o mini by 2.376.874.5+2.3
- simple_safety_testsQwen2.5 Instruct 72B by 2.297.8100+2.2
- XSTestQwen2.5 Instruct 72B by 1.99697.9+1.9
- anthropic_red_teamQwen2.5 Instruct 72B by 1.398.399.6+1.3
- AA IntelligenceQwen2.5 Instruct 72B by 16.77.7+1
- SimpleQAtie9.99.1tie
- HumanEvaltie87.286.6tie
- Artificial Analysis Coding Indextie11.411.9tie
Questions people ask
Which is better, GPT-4o mini or Qwen2.5 Instruct 72B?
Qwen2.5 Instruct 72B wins all three areas where both have results: coding, reasoning and following instructions. GPT-4o mini wins none, but costs 23% less.
Which is better for coding?
Qwen2.5 Instruct 72B. It wins 2 of the 2 coding tests both models report; GPT-4o mini wins none.
Which is cheaper?
GPT-4o mini costs $0.15 per million input tokens and $0.60 per million output tokens; Qwen2.5 Instruct 72B costs $0.47 and $0.49. That makes GPT-4o mini about 23% cheaper for the same work.
How do you compare the two?
We use the 25 benchmark tests both models have published scores on. The verdict counts the 5 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 20 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.