Model comparison
Llama 3-8B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 25.3 on the Noometry Index. Llama 3-8B costs 8.0× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Last verified . 19 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Llama 3-8B 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 7.8.
- The biggest single-benchmark swing is DTBench: 43.9% for Llama 3-8B and 88% for Qwen3.8 27B.
- Llama 3-8B is cheaper at $0.14 / $0.14 per million input/output tokens, against $0.99 / $1.49 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 8K.
Side by side
| Llama 3-8B | Qwen3.8 27B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 25.3 | 45.3 |
| Released | 2024-04-18 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 8K | 262K |
| Max output | 8K | 33K |
| Input $ / M tokens | $0.14 | $0.99 |
| Output $ / M tokens | $0.14 | $1.49 |
| Results tracked | 34 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Llama 3-8B: 30.5 (#285), Qwen3.8 27B: 50.3 (#42)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1152 | 1487 |
| LMArena WebDev | — | 1592 |
| SciCode | — | 46.6% |
| BigCodeBench Instruct | 31.9% | — |
| BigCodeBench Complete | 36.9% | — |
| HumanEval+ | 56.7% | — |
| MBPP+ | 54.8% | — |
Reasoning Qwen3.8 27B leads
Llama 3-8B: 13.8 (#321), Qwen3.8 27B: 31.5 (#71)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1133 | 1459 |
| DTBench | 43.9% | 88% |
| Epoch Capabilities Index | 116.33 | 149.38 |
| NYT Connections (extended) | — | 54.5% |
| CritPt | — | 5.4% |
| Chess Puzzles | 0% | — |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| Adversarial NLI | 57.3% | — |
| ForecastBench | 58.6 | — |
| WinoGrande | 75.7% | — |
Math Qwen3.8 27B leads
Llama 3-8B: 8.3 (#317), Qwen3.8 27B: 35.7 (#139)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1151 | 1453 |
| OTIS Mock AIME 2024-2025 | 1.9% | — |
| ProofBench | — | 16% |
| MATH Level 5 | 6.1% | — |
Knowledge Qwen3.8 27B leads
Llama 3-8B: 7.8 (#303), Qwen3.8 27B: 40.9 (#105)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1113 | 1482 |
| GPQA Diamond | 26.1% | — |
| ARC (AI2) Challenge | 82.8% | — |
| MMLU | 68.8% | — |
| OpenBookQA | 82.6% | — |
| TriviaQA | 67.7% | — |
Multimodal Not comparable
Llama 3-8B: —, Qwen3.8 27B: 40.5 (#36)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1270 |
Multilingual Qwen3.8 27B leads
Llama 3-8B: 31.0 (#256), Qwen3.8 27B: 54.0 (#54)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1098 | 1431 |
| LMArena Chinese | 1076 | 1504 |
| LMArena French | 1158 | 1468 |
| LMArena German | 1104 | 1435 |
| LMArena Japanese | 967 | 1392 |
| LMArena Korean | 1003 | 1389 |
| LMArena Russian | 1109 | 1421 |
| LMArena Spanish | 1172 | 1444 |
Instruction Following Qwen3.8 27B leads
Llama 3-8B: 58.7 (#255), Qwen3.8 27B: 75.8 (#54)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1127 | 1436 |
Long Context Qwen3.8 27B leads
Llama 3-8B: 34.2 (#246), Qwen3.8 27B: 44.3 (#66)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1128 | 1448 |
Writing & Preference Qwen3.8 27B leads
Llama 3-8B: 37.7 (#251), Qwen3.8 27B: 65.9 (#41)
| Benchmark | Llama 3-8B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1166 | 1441 |
| LMArena Creative Writing | 1150 | 1382 |
| LMArena Multi-Turn | 1152 | 1443 |
| EQ-Bench Creative Writing | — | 1671 |
Frequently asked questions
Is Llama 3-8B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 45.3 to 25.3 on the Noometry Index. Llama 3-8B costs 8.0× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Which is cheaper, Llama 3-8B or Qwen3.8 27B?
Llama 3-8B is cheaper. It lists at $0.14 per million input tokens and $0.14 per million output tokens; Qwen3.8 27B lists at $0.99 and $1.49.
Is Llama 3-8B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.3 versus 30.5 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 8K.
How many benchmarks do Llama 3-8B and Qwen3.8 27B share?
19 benchmarks have published results for both models. Llama 3-8B has 34 scored results on Noometry and Qwen3.8 27B has 28.