Model comparison
Llama 3.1-70B vs QwQ-32B
QwQ-32B is the stronger model overall, scoring 39.4 to 29.4 on the Noometry Index.
Last verified . 23 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 23 benchmarks with published results for both. Llama 3.1-70B 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 36.2 to 12.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 3.6% for Llama 3.1-70B and 59.2% for QwQ-32B.
Side by side
| Llama 3.1-70B | QwQ-32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 29.4 | 39.4 |
| Released | 2024-07-23 | 2024-11-28 |
| Weights | Open | Open |
| Context window | 128K | — |
| Max output | 4K | — |
| Input $ / M tokens | $0.40 | — |
| Output $ / M tokens | $0.40 | — |
| Results tracked | 35 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding QwQ-32B leads
Llama 3.1-70B: 29.8 (#292), QwQ-32B: 34.9 (#220)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| BigCodeBench Instruct | 46.1% | 44.6% |
| LMArena Coding | 1260 | 1333 |
| BigCodeBench Complete | 54.8% | 54.4% |
| Aider Polyglot | — | 20.9% |
| WeirdML | 9% | — |
| LiveBench Coding | — | 72.2% |
Agentic & Tool Use Not comparable
Llama 3.1-70B: 24.9 (#104), QwQ-32B: —
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| TheAgentCompany | 6.9% | — |
| BALROG | 27.9% | — |
Reasoning QwQ-32B leads
Llama 3.1-70B: 20.8 (#215), QwQ-32B: 22.9 (#171)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1326 |
| Epoch Capabilities Index | 125.91 | 137.6 |
| Chess Puzzles | — | 5% |
| LiveBench Reasoning | — | 83.5% |
| DTBench | 60% | — |
| LiveBench Data Analysis | — | 65% |
| LMCA | 14.8% | — |
| ForecastBench | — | 58.3 |
| LiveBench | — | 72% |
Math QwQ-32B leads
Llama 3.1-70B: 12.7 (#298), QwQ-32B: 36.2 (#123)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 3.6% | 59.2% |
| LMArena Math | 1252 | 1359 |
| Omni-MATH | 21% | — |
| LiveBench Math | — | 77.8% |
| MATH Level 5 | 36.7% | — |
Knowledge QwQ-32B leads
Llama 3.1-70B: 24.1 (#264), QwQ-32B: 37.2 (#150)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| GPQA Diamond | 44.2% | 65.3% |
| LMArena Expert | 1209 | 1323 |
| MMLU-Pro | 65.3% | — |
| Confabulations | — | 15.6% |
| GPQA (HELM) | 42.6% | — |
| MMLU | 80.1% | — |
Multilingual QwQ-32B leads
Llama 3.1-70B: 38.9 (#220), QwQ-32B: 45.0 (#171)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1219 | 1306 |
| LMArena Chinese | 1215 | 1378 |
| LMArena French | 1260 | 1337 |
| LMArena German | 1222 | 1313 |
| LMArena Japanese | 1132 | 1263 |
| LMArena Korean | 1140 | 1278 |
| LMArena Russian | 1234 | 1297 |
| LMArena Spanish | 1253 | 1353 |
Instruction Following QwQ-32B leads
Llama 3.1-70B: 65.5 (#218), QwQ-32B: 72.8 (#131)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| LMArena Instruction Following | 1231 | 1297 |
| LiveBench Instruction Following | — | 81.8% |
| IFEval | 82.1% | — |
Long Context QwQ-32B leads
Llama 3.1-70B: 37.6 (#208), QwQ-32B: 49.0 (#11)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| LMArena Longer Query | 1241 | 1308 |
| Fiction.LiveBench | — | 83.3% |
Writing & Preference QwQ-32B leads
Llama 3.1-70B: 35.6 (#262), QwQ-32B: 50.9 (#175)
| Benchmark | Llama 3.1-70B | QwQ-32B |
|---|---|---|
| LMArena Text | 1261 | 1329 |
| LMArena Creative Writing | 1232 | 1288 |
| EQ-Bench Creative Writing | 784 | 1257 |
| LMArena Multi-Turn | 1256 | 1314 |
| Short-Story Creative Writing | — | 80.2% |
| WildBench | 75.8% | — |
| LiveBench Language | — | 51.4% |
Frequently asked questions
Is Llama 3.1-70B better than QwQ-32B?
QwQ-32B is the stronger model overall, scoring 39.4 to 29.4 on the Noometry Index.
Is Llama 3.1-70B or QwQ-32B better for coding?
QwQ-32B scores higher on coding benchmarks: 34.9 versus 29.8 in the Noometry coding category.
How many benchmarks do Llama 3.1-70B and QwQ-32B share?
23 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and QwQ-32B has 36.