Model comparison
Llama 2-70B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 24.1 on the Noometry Index.
Last verified . 19 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Llama 2-70B scores higher in 0 categories and Qwen3.8 27B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.8 27B leads 40.9 to 7.4.
- The biggest single-benchmark swing is DTBench: 41.6% for Llama 2-70B and 88% for Qwen3.8 27B.
Side by side
| Llama 2-70B | Qwen3.8 27B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 24.1 | 45.3 |
| Released | 2023-07-18 | 2026-08-14 |
| Weights | Open | Open |
| Context window | — | 262K |
| Max output | — | 33K |
| Input $ / M tokens | — | $0.99 |
| Output $ / M tokens | — | $1.49 |
| Results tracked | 35 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Llama 2-70B: 30.8 (#281), Qwen3.8 27B: 50.3 (#42)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1079 | 1487 |
| LMArena WebDev | — | 1592 |
| SciCode | — | 46.6% |
Reasoning Qwen3.8 27B leads
Llama 2-70B: 13.7 (#324), Qwen3.8 27B: 31.5 (#71)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1073 | 1459 |
| DTBench | 41.6% | 88% |
| Epoch Capabilities Index | 113.59 | 149.38 |
| NYT Connections (extended) | — | 54.5% |
| CritPt | — | 5.4% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| BIG-Bench Hard | 64.9% | — |
| CommonsenseQA 2.0 | 50% | — |
| ForecastBench | 51.4 | — |
| HellaSwag | 85.3% | — |
| LAMBADA | 78.9% | — |
| PIQA | 82.8% | — |
| WinoGrande | 80.2% | — |
Math Qwen3.8 27B leads
Llama 2-70B: 7.5 (#320), Qwen3.8 27B: 35.7 (#139)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1091 | 1453 |
| OTIS Mock AIME 2024-2025 | 0% | — |
| ProofBench | — | 16% |
| MATH Level 5 | 3.3% | — |
| GSM8K | 69.6% | — |
Knowledge Qwen3.8 27B leads
Llama 2-70B: 7.4 (#305), Qwen3.8 27B: 40.9 (#105)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1039 | 1482 |
| GPQA Diamond | 26.3% | — |
| ARC (AI2) Challenge | 78.3% | — |
| BoolQ | 88.6% | — |
| MMLU | 69.9% | — |
| OpenBookQA | 60.2% | — |
| TriviaQA | 87.6% | — |
Multimodal Not comparable
Llama 2-70B: —, Qwen3.8 27B: 40.5 (#36)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1270 |
Multilingual Qwen3.8 27B leads
Llama 2-70B: 27.9 (#269), Qwen3.8 27B: 54.0 (#54)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1045 | 1431 |
| LMArena Chinese | 994 | 1504 |
| LMArena French | 1089 | 1468 |
| LMArena German | 1041 | 1435 |
| LMArena Japanese | 927 | 1392 |
| LMArena Korean | 964 | 1389 |
| LMArena Russian | 1083 | 1421 |
| LMArena Spanish | 1142 | 1444 |
Instruction Following Qwen3.8 27B leads
Llama 2-70B: 55.1 (#273), Qwen3.8 27B: 75.8 (#54)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1071 | 1436 |
Long Context Qwen3.8 27B leads
Llama 2-70B: 32.2 (#265), Qwen3.8 27B: 44.3 (#66)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1062 | 1448 |
Writing & Preference Qwen3.8 27B leads
Llama 2-70B: 32.5 (#274), Qwen3.8 27B: 65.9 (#41)
| Benchmark | Llama 2-70B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1115 | 1441 |
| LMArena Creative Writing | 1075 | 1382 |
| LMArena Multi-Turn | 1088 | 1443 |
| EQ-Bench Creative Writing | — | 1671 |
Frequently asked questions
Is Llama 2-70B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 24.1 on the Noometry Index.
Is Llama 2-70B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.3 versus 30.8 in the Noometry coding category.
How many benchmarks do Llama 2-70B and Qwen3.8 27B share?
19 benchmarks have published results for both models. Llama 2-70B has 35 scored results on Noometry and Qwen3.8 27B has 28.