Model comparison
Llama 3-70B vs QwQ-32B
QwQ-32B is the stronger model overall, scoring 39.8 to 28.8 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Llama 3-70B scores higher in 1 category and QwQ-32B in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where QwQ-32B leads 38.0 to 12.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 4.3% for Llama 3-70B and 59.2% for QwQ-32B.
Side by side
| Llama 3-70B | QwQ-32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.8 | 39.8 |
| Released | 2024-04-18 | 2024-11-28 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 31 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Llama 3-70B: 35.8 (#218), QwQ-32B: 35.4 (#226)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| BigCodeBench Instruct | 43.6% | 44.6% |
| LMArena Coding | 1206 | 1333 |
| BigCodeBench Complete | 54.5% | 54.4% |
| Aider Polyglot | — | 20.9% |
| LiveBench Coding | — | 72.2% |
| HumanEval+ | 72% | — |
| MBPP+ | 69% | — |
Agentic & Tool Use Not comparable
Llama 3-70B: 21.1 (#139), QwQ-32B: —
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| Cybench | 5% | — |
Reasoning QwQ-32B leads
Llama 3-70B: 18.0 (#288), QwQ-32B: 23.7 (#174)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| LMArena Hard Prompts | 1195 | 1325 |
| Epoch Capabilities Index | 122.93 | 137.6 |
| ForecastBench | 57.1 | 58.3 |
| Kagi LLM Benchmark | 35.1% | — |
| Chess Puzzles | — | 5% |
| LiveBench Reasoning | — | 83.5% |
| DTBench | 54.2% | — |
| LiveBench Data Analysis | — | 65% |
| LiveBench | — | 72% |
| WinoGrande | 83.5% | — |
Math QwQ-32B leads
Llama 3-70B: 12.8 (#305), QwQ-32B: 38.0 (#143)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 4.3% | 59.2% |
| LMArena Math | 1218 | 1359 |
| LiveBench Math | — | 77.8% |
| MATH Level 5 | 22.6% | — |
Knowledge QwQ-32B leads
Llama 3-70B: 20.8 (#277), QwQ-32B: 37.2 (#158)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| GPQA Diamond | 40.6% | 65.3% |
| LMArena Expert | 1149 | 1324 |
| Confabulations | — | 15.6% |
| MMLU | 79.3% | — |
Multilingual QwQ-32B leads
Llama 3-70B: 33.6 (#251), QwQ-32B: 44.8 (#176)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1142 | 1305 |
| LMArena Chinese | 1114 | 1378 |
| LMArena French | 1232 | 1336 |
| LMArena German | 1169 | 1313 |
| LMArena Japanese | 1017 | 1262 |
| LMArena Korean | 1017 | 1279 |
| LMArena Russian | 1159 | 1297 |
| LMArena Spanish | 1241 | 1354 |
Instruction Following QwQ-32B leads
Llama 3-70B: 62.5 (#238), QwQ-32B: 72.6 (#137)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| LMArena Instruction Following | 1194 | 1297 |
| LiveBench Instruction Following | — | 81.8% |
Long Context QwQ-32B leads
Llama 3-70B: 35.6 (#240), QwQ-32B: 49.0 (#11)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| LMArena Longer Query | 1174 | 1308 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference QwQ-32B leads
Llama 3-70B: 42.8 (#231), QwQ-32B: 50.6 (#180)
| Benchmark | Llama 3-70B | QwQ-32B |
|---|---|---|
| LMArena Text | 1221 | 1329 |
| LMArena Creative Writing | 1210 | 1288 |
| LMArena Multi-Turn | 1223 | 1314 |
| Short-Story Creative Writing | — | 80.2% |
| EQ-Bench Creative Writing | — | 1257 |
| LiveBench Language | — | 51.4% |
Frequently asked questions
Is Llama 3-70B better than QwQ-32B?
QwQ-32B is the stronger model overall, scoring 39.8 to 28.8 on the Noometry Index.
Is Llama 3-70B or QwQ-32B better for coding?
They score almost the same on coding (35.8 vs 35.4); test both on your own repository before choosing.
How many benchmarks do Llama 3-70B and QwQ-32B share?
23 benchmarks have published results for both models. Llama 3-70B has 31 scored results on Noometry and QwQ-32B has 36.