Model comparison
Phi-4 vs QwQ-32B
QwQ-32B is the stronger model overall, scoring 39.8 to 31.2 on the Noometry Index.
Last verified . 32 shared benchmarks.
Microsoft
Alibaba (Qwen)Summary
- They share 32 benchmarks with published results for both. Phi-4 scores higher in 0 categories and QwQ-32B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where QwQ-32B leads 38.0 to 20.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 13.8% for Phi-4 and 59.2% for QwQ-32B.
Side by side
| Phi-4 | QwQ-32B | |
|---|---|---|
| Provider | Microsoft | Alibaba (Qwen) |
| Noometry Index | 31.2 | 39.8 |
| Released | 2024-12-11 | 2024-11-28 |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.07 | — |
| Output $ / M tokens | $0.14 | — |
| Results tracked | 37 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding QwQ-32B leads
Phi-4: 34.4 (#239), QwQ-32B: 35.4 (#226)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| BigCodeBench Instruct | 45.5% | 44.6% |
| LiveBench Coding | 30.7% | 72.2% |
| LMArena Coding | 1231 | 1333 |
| BigCodeBench Complete | 55.4% | 54.4% |
| Aider Polyglot | — | 20.9% |
Agentic & Tool Use Not comparable
Phi-4: 22.8 (#128), QwQ-32B: —
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 28.8% | — |
| BALROG | 11.6% | — |
Reasoning QwQ-32B leads
Phi-4: 17.7 (#291), QwQ-32B: 23.7 (#174)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| Chess Puzzles | 1% | 5% |
| LiveBench Reasoning | 47.8% | 83.5% |
| LMArena Hard Prompts | 1220 | 1325 |
| LiveBench Data Analysis | 45.2% | 65% |
| Epoch Capabilities Index | 130.42 | 137.6 |
| LiveBench | 41.6% | 72% |
| ForecastBench | — | 58.3 |
Math QwQ-32B leads
Phi-4: 20.8 (#285), QwQ-32B: 38.0 (#143)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 13.8% | 59.2% |
| LiveBench Math | 42% | 77.8% |
| LMArena Math | 1246 | 1359 |
| MATH Level 5 | 64.9% | — |
Knowledge QwQ-32B leads
Phi-4: 32.6 (#209), QwQ-32B: 37.2 (#158)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| GPQA Diamond | 56.1% | 65.3% |
| Confabulations | 29.4% | 15.6% |
| LMArena Expert | 1203 | 1324 |
| Vectara Hallucination Rate | 3.7% | — |
| MMLU | 84.8% | — |
Multilingual QwQ-32B leads
Phi-4: 37.2 (#237), QwQ-32B: 44.8 (#176)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1197 | 1305 |
| LMArena Chinese | 1212 | 1378 |
| LMArena French | 1224 | 1336 |
| LMArena German | 1222 | 1313 |
| LMArena Japanese | 1158 | 1262 |
| LMArena Korean | 1151 | 1279 |
| LMArena Russian | 1209 | 1297 |
| LMArena Spanish | 1234 | 1354 |
Instruction Following QwQ-32B leads
Phi-4: 60.4 (#251), QwQ-32B: 72.6 (#137)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| LiveBench Instruction Following | 58.4% | 81.8% |
| LMArena Instruction Following | 1201 | 1297 |
Long Context QwQ-32B leads
Phi-4: 36.9 (#226), QwQ-32B: 49.0 (#11)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| LMArena Longer Query | 1217 | 1308 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference QwQ-32B leads
Phi-4: 40.5 (#244), QwQ-32B: 50.6 (#180)
| Benchmark | Phi-4 | QwQ-32B |
|---|---|---|
| LMArena Text | 1217 | 1329 |
| LMArena Creative Writing | 1182 | 1288 |
| Short-Story Creative Writing | 62.6% | 80.2% |
| LMArena Multi-Turn | 1206 | 1314 |
| LiveBench Language | 25.6% | 51.4% |
| EQ-Bench Creative Writing | — | 1257 |
Frequently asked questions
Is Phi-4 better than QwQ-32B?
QwQ-32B is the stronger model overall, scoring 39.8 to 31.2 on the Noometry Index.
Is Phi-4 or QwQ-32B better for coding?
QwQ-32B scores higher on coding benchmarks: 35.4 versus 34.4 in the Noometry coding category.
How many benchmarks do Phi-4 and QwQ-32B share?
32 benchmarks have published results for both models. Phi-4 has 37 scored results on Noometry and QwQ-32B has 36.