Llama 3.2 3B Instruct vs Qwen3.6 Max
Meta vs Alibaba — benchmarks, pricing, and capabilities side by side.
- •Qwen3.6 Max has the higher intelligence index (88.8 vs 30.8)
- •Llama 3.2 3B Instruct is cheaper ($0.05 vs $1.04 per 1M input)
- •Llama 3.2 3B Instruct is faster
- •Qwen3.6 Max has a larger context window (262K)
| Llama 3.2 3B Instruct | Qwen3.6 Max | |
|---|---|---|
| Intelligence index | 30.8 | 88.8 |
| Developer | Meta | Alibaba |
| Type | LLM | LLM |
| Access | Open weights | API only |
| Context window | 131,072 tokens | 262,144 tokens |
| Input price | $0.05 / 1M | $1.04 / 1M |
| Output price | $0.34 / 1M | $6.24 / 1M |
| Speed | 172 tok/s | 36 tok/s |
| Released | September 25, 2024 | April 27, 2026 |
| Parameters | — | — |
| Input modalities | Text | Text |
| Output modalities | Text | Text |
Shared benchmarks
Llama 3.2 3B Instruct
Qwen3.6 Max
GPQA Diamond
32.8
88.8
Humanity’s Last Exam
5.2
28.9