Model comparison
Llama 3.2 3B vs Qwen3.8 27B
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 28.9 on the Noometry Index. Llama 3.2 3B costs 9.3× less per token, which makes it the better buy when Qwen3.8 27B's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 14 benchmarks with published results for both. Llama 3.2 3B scores higher in 0 categories and Qwen3.8 27B in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen3.8 27B leads 65.8 to 24.7.
- Llama 3.2 3B is cheaper at $0.05 / $0.33 per million input/output tokens, against $0.99 / $1.49 for Qwen3.8 27B.
- Qwen3.8 27B accepts more context: 262K tokens versus 131K.
Side by side
| Llama 3.2 3B | Qwen3.8 27B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 28.9 | 46.0 |
| Released | 2024-09-24 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 118K | 33K |
| Input $ / M tokens | $0.05 | $0.99 |
| Output $ / M tokens | $0.33 | $1.49 |
| Results tracked | 18 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 27B leads
Llama 3.2 3B: 27.6 (#319), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Coding | 1098 | 1482 |
| LMArena WebDev | — | 1593 |
| SciCode | — | 46.6% |
| BigCodeBench Instruct | 23.4% | — |
| BigCodeBench Complete | 28.3% | — |
Agentic & Tool Use Qwen3.8 27B leads
Llama 3.2 3B: 20.1 (#143), Qwen3.8 27B: 32.9 (#57)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| Berkeley Function Calling Leaderboard | 21.9% | — |
| BALROG | 10.1% | — |
Reasoning Qwen3.8 27B leads
Llama 3.2 3B: 21.0 (#228), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Hard Prompts | 1095 | 1460 |
| ARC-AGI-2 | — | 42.4% |
| NYT Connections (extended) | — | 54.5% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 5.4% |
| DTBench | — | 88% |
| LMCA | — | 41.4% |
| Surface Evolver Bench | — | 45% |
| Epoch Capabilities Index | — | 149.38 |
Math Qwen3.8 27B leads
Llama 3.2 3B: 32.4 (#214), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Math | 1126 | 1456 |
| ProofBench | — | 16% |
Knowledge Qwen3.8 27B leads
Llama 3.2 3B: 29.7 (#235), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1090 | 1482 |
Multimodal Not comparable
Llama 3.2 3B: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.8 27B leads
Llama 3.2 3B: 26.2 (#281), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1019 | 1430 |
| LMArena Chinese | 1017 | 1504 |
| LMArena German | 1056 | 1438 |
| LMArena Russian | 949 | 1415 |
| LMArena French | — | 1465 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Qwen3.8 27B leads
Llama 3.2 3B: 56.0 (#275), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1089 | 1439 |
Long Context Qwen3.8 27B leads
Llama 3.2 3B: 33.4 (#261), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1100 | 1450 |
Writing & Preference Qwen3.8 27B leads
Llama 3.2 3B: 24.7 (#307), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Llama 3.2 3B | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1110 | 1441 |
| LMArena Creative Writing | 1094 | 1384 |
| EQ-Bench Creative Writing | 595 | 1671 |
| LMArena Multi-Turn | 1105 | 1441 |
Frequently asked questions
Is Llama 3.2 3B better than Qwen3.8 27B?
Qwen3.8 27B is the stronger model overall, scoring 46.0 to 28.9 on the Noometry Index. Llama 3.2 3B costs 9.3× 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 3.2 3B or Qwen3.8 27B?
Llama 3.2 3B is cheaper. It lists at $0.05 per million input tokens and $0.33 per million output tokens; Qwen3.8 27B lists at $0.99 and $1.49.
Is Llama 3.2 3B or Qwen3.8 27B better for coding?
Qwen3.8 27B scores higher on coding benchmarks: 50.5 versus 27.6 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 27B does, with 262K tokens against 131K.
How many benchmarks do Llama 3.2 3B and Qwen3.8 27B share?
14 benchmarks have published results for both models. Llama 3.2 3B has 18 scored results on Noometry and Qwen3.8 27B has 31.