Qwen3 14B — Specifications
| Desenvolvedor | Alibaba |
|---|---|
| Tipo | LLM (dense) |
| Modalidade | Text → Text |
| Parâmetros | 14B |
| Janela de contexto | 128K |
| Saída máxima | — |
| Licença | Apache 2.0 (open) |
| Pesos abertos | ✅ Yes |
| Lançamento | 2025 |
| Input price | $0.12 /1M |
| Output price | $0.24 /1M |
| API providers | Alibaba, OpenRouter, Ollama |
🖥️ Run it locally
| VRAM (4-bit) | ~9 GB |
|---|---|
| Minimum GPU | RTX 4070 12GB (Q4) |
A mid-size dense Qwen3 (Apache 2.0, 128K context) that hits a sweet spot for local use — runs 4-bit on a 12GB card while staying genuinely capable.