# Llama 2-70B vs QwQ-32B

> QwQ-32B is the stronger model overall, scoring 39.4 to 24.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/llama-2-70b-vs-qwq-32b
- Last updated: 2026-10-10
- Shared benchmarks: 21

## Summary

- They share 21 benchmarks with published results for both. Llama 2-70B scores higher in 0 categories and QwQ-32B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where QwQ-32B leads 37.2 to 7.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 0% for Llama 2-70B and 59.2% for QwQ-32B.

## Snapshot

| | Llama 2-70B | QwQ-32B |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 24.1 | 39.4 |
| Rank | 343 | 153 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Open | Open |

## Coding

- Llama 2-70B: 30.8 (#281)
- QwQ-32B: 34.9 (#220)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Coding | 1079 | 1333 |
| Aider Polyglot | — | 20.9% |
| BigCodeBench Instruct | — | 44.6% |
| LiveBench Coding | — | 72.2% |
| BigCodeBench Complete | — | 54.4% |

## Reasoning

- Llama 2-70B: 13.7 (#324)
- QwQ-32B: 22.9 (#171)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1326 |
| Epoch Capabilities Index | 113.59 | 137.6 |
| ForecastBench | 51.4 | 58.3 |
| Chess Puzzles | — | 5% |
| LiveBench Reasoning | — | 83.5% |
| DTBench | 41.6% | — |
| LiveBench Data Analysis | — | 65% |
| BIG-Bench Hard | 64.9% | — |
| CommonsenseQA 2.0 | 50% | — |
| HellaSwag | 85.3% | — |
| LAMBADA | 78.9% | — |
| LiveBench | — | 72% |
| PIQA | 82.8% | — |
| WinoGrande | 80.2% | — |

## Math

- Llama 2-70B: 7.5 (#320)
- QwQ-32B: 36.2 (#123)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 0% | 59.2% |
| LMArena Math | 1091 | 1359 |
| LiveBench Math | — | 77.8% |
| MATH Level 5 | 3.3% | — |
| GSM8K | 69.6% | — |

## Knowledge

- Llama 2-70B: 7.4 (#305)
- QwQ-32B: 37.2 (#150)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| GPQA Diamond | 26.3% | 65.3% |
| LMArena Expert | 1039 | 1323 |
| Confabulations | — | 15.6% |
| ARC (AI2) Challenge | 78.3% | — |
| BoolQ | 88.6% | — |
| MMLU | 69.9% | — |
| OpenBookQA | 60.2% | — |
| TriviaQA | 87.6% | — |

## Multilingual

- Llama 2-70B: 27.9 (#269)
- QwQ-32B: 45.0 (#171)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Non-English | 1045 | 1306 |
| LMArena Chinese | 994 | 1378 |
| LMArena French | 1089 | 1337 |
| LMArena German | 1041 | 1313 |
| LMArena Japanese | 927 | 1263 |
| LMArena Korean | 964 | 1278 |
| LMArena Russian | 1083 | 1297 |
| LMArena Spanish | 1142 | 1353 |

## Instruction Following

- Llama 2-70B: 55.1 (#273)
- QwQ-32B: 72.8 (#131)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1297 |
| LiveBench Instruction Following | — | 81.8% |

## Long Context

- Llama 2-70B: 32.2 (#265)
- QwQ-32B: 49.0 (#11)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Longer Query | 1062 | 1308 |
| Fiction.LiveBench | — | 83.3% |

## Writing & Preference

- Llama 2-70B: 32.5 (#274)
- QwQ-32B: 50.9 (#175)

| Benchmark | Llama 2-70B | QwQ-32B |
|---|---|---|
| LMArena Text | 1115 | 1329 |
| LMArena Creative Writing | 1075 | 1288 |
| LMArena Multi-Turn | 1088 | 1314 |
| Short-Story Creative Writing | — | 80.2% |
| EQ-Bench Creative Writing | — | 1257 |
| LiveBench Language | — | 51.4% |

## FAQ

### Is Llama 2-70B better than QwQ-32B?

QwQ-32B is the stronger model overall, scoring 39.4 to 24.1 on the Noometry Index.

### Is Llama 2-70B or QwQ-32B better for coding?

QwQ-32B scores higher on coding benchmarks: 34.9 versus 30.8 in the Noometry coding category.

### How many benchmarks do Llama 2-70B and QwQ-32B share?

21 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and QwQ-32B has 36.