Model comparison
Llama 3-8B vs Qwen3.5 27B
Qwen3.5 27B is the stronger model overall, scoring 41.4 to 25.3 on the Noometry Index. Llama 3-8B costs 5.9× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Llama 3-8B scores higher in 0 categories and Qwen3.5 27B in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen3.5 27B leads 37.4 to 7.8.
- The biggest single-benchmark swing is DTBench: 43.9% for Llama 3-8B and 82.4% for Qwen3.5 27B.
- Llama 3-8B is cheaper at $0.14 / $0.14 per million input/output tokens, against $0.30 / $2.40 for Qwen3.5 27B.
- Qwen3.5 27B accepts more context: 262K tokens versus 8K.
Side by side
| Llama 3-8B | Qwen3.5 27B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 25.3 | 41.4 |
| Released | 2024-04-18 | 2026-02-23 |
| Weights | Open | Open |
| Context window | 8K | 262K |
| Max output | 8K | 66K |
| Input $ / M tokens | $0.14 | $0.30 |
| Output $ / M tokens | $0.14 | $2.40 |
| Results tracked | 34 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.5 27B leads
Llama 3-8B: 30.5 (#285), Qwen3.5 27B: 38.6 (#159)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Coding | 1152 | 1427 |
| LMArena WebDev | — | 1357 |
| WeirdML | — | 39.5% |
| BigCodeBench Instruct | 31.9% | — |
| BigCodeBench Complete | 36.9% | — |
| ALE-Bench | — | 349.45 |
| HumanEval+ | 56.7% | — |
| MBPP+ | 54.8% | — |
Agentic & Tool Use Not comparable
Llama 3-8B: —, Qwen3.5 27B: —
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| Vending-Bench 2 | — | 201.98 |
Reasoning Qwen3.5 27B leads
Llama 3-8B: 13.8 (#321), Qwen3.5 27B: 26.6 (#116)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Hard Prompts | 1133 | 1414 |
| DTBench | 43.9% | 82.4% |
| NYT Connections (extended) | — | 47.9% |
| Chess Puzzles | 0% | — |
| Thematic Generalization | — | 45.5% |
| LMCA | — | 34% |
| Adversarial NLI | 57.3% | — |
| Epoch Capabilities Index | 116.33 | — |
| ForecastBench | 58.6 | — |
| WinoGrande | 75.7% | — |
Math Qwen3.5 27B leads
Llama 3-8B: 8.3 (#317), Qwen3.5 27B: 36.8 (#103)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Math | 1151 | 1428 |
| MathArena Final-Answer Competitions | — | 56.7% |
| OTIS Mock AIME 2024-2025 | 1.9% | — |
| MATH Level 5 | 6.1% | — |
Knowledge Qwen3.5 27B leads
Llama 3-8B: 7.8 (#303), Qwen3.5 27B: 37.4 (#149)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Expert | 1113 | 1428 |
| GPQA Diamond | 26.1% | — |
| Vectara Hallucination Rate | — | 12.1% |
| ARC (AI2) Challenge | 82.8% | — |
| MMLU | 68.8% | — |
| OpenBookQA | 82.6% | — |
| TriviaQA | 67.7% | — |
Multimodal Not comparable
Llama 3-8B: —, Qwen3.5 27B: 38.7 (#56)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Vision | — | 1240 |
Multilingual Qwen3.5 27B leads
Llama 3-8B: 31.0 (#256), Qwen3.5 27B: 51.0 (#112)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Non-English | 1098 | 1390 |
| LMArena Chinese | 1076 | 1478 |
| LMArena French | 1158 | 1409 |
| LMArena German | 1104 | 1393 |
| LMArena Japanese | 967 | 1345 |
| LMArena Korean | 1003 | 1359 |
| LMArena Russian | 1109 | 1389 |
| LMArena Spanish | 1172 | 1408 |
Instruction Following Qwen3.5 27B leads
Llama 3-8B: 58.7 (#255), Qwen3.5 27B: 73.7 (#114)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Instruction Following | 1127 | 1393 |
Long Context Qwen3.5 27B leads
Llama 3-8B: 34.2 (#246), Qwen3.5 27B: 43.1 (#103)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Longer Query | 1128 | 1413 |
Writing & Preference Qwen3.5 27B leads
Llama 3-8B: 37.7 (#251), Qwen3.5 27B: 59.5 (#106)
| Benchmark | Llama 3-8B | Qwen3.5 27B |
|---|---|---|
| LMArena Text | 1166 | 1409 |
| LMArena Creative Writing | 1150 | 1362 |
| LMArena Multi-Turn | 1152 | 1410 |
Frequently asked questions
Is Llama 3-8B better than Qwen3.5 27B?
Qwen3.5 27B is the stronger model overall, scoring 41.4 to 25.3 on the Noometry Index. Llama 3-8B costs 5.9× less per token, which makes it the better buy when Qwen3.5 27B's lead doesn't matter for your workload.
Which is cheaper, Llama 3-8B or Qwen3.5 27B?
Llama 3-8B is cheaper. It lists at $0.14 per million input tokens and $0.14 per million output tokens; Qwen3.5 27B lists at $0.30 and $2.40.
Is Llama 3-8B or Qwen3.5 27B better for coding?
Qwen3.5 27B scores higher on coding benchmarks: 38.6 versus 30.5 in the Noometry coding category.
Which has the bigger context window?
Qwen3.5 27B does, with 262K tokens against 8K.
How many benchmarks do Llama 3-8B and Qwen3.5 27B share?
18 benchmarks have published results for both models. Llama 3-8B has 34 scored results on Noometry and Qwen3.5 27B has 28.