MODEL / ALIBABA QWEN

Qwen3 4B

Thinking mode can increase latency and output length.

总参数4B

估算内存~3.6 GBGGUF Q4

上下文32K+Text

LicenseApache-2.0100+ languages and dialects

01 / 适合

02 / 估算适配

03 / 部署方式

OllamaEasyollama run qwen3:4b

llama.cppMediumllama-server -m /path/to/verified-model.gguf --port 8080

LM StudioEasySearch “Qwen/Qwen3-4B” → Load a Q4/INT4 build → Start Server

MLX LMMediummlx_lm.chat --model Qwen/Qwen3-4B

注意

Thinking mode can increase latency and output length.