← CanIRunLocal · All hardware · All models

Can I run Qwen3 1.7B locally?

Alibaba · 2.03B parameters · chat, reasoning.

Ollama weight files: Q4_K_M 1.4 GB (qwen3:1.7b); Q8_0 2.2 GB (qwen3:1.7b-q8_0).

Yes on 62 machines in this list. The smallest is M1 8GB: Q8_0, about 2.9 GB, roughly 25 tokens per second, via Ollama. ollama run qwen3:1.7b-q8_0

DGX Spark 128GB: Q8_0, ~2.9 GB, fits in fast memory, ~93 tok/s — ollama run qwen3:1.7b-q8_0
Mac Studio M5 Max 128GB: Q8_0, ~2.9 GB, fits in fast memory, ~223 tok/s — ollama run qwen3:1.7b-q8_0
Mac Studio M5 Ultra 256GB: Q8_0, ~2.9 GB, fits in fast memory, ~436 tok/s — ollama run qwen3:1.7b-q8_0

Setup on DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte)

Ollama: ollama run qwen3:1.7b-q8_0
LM Studio: lms get unsloth/Qwen3-1.7B-GGUF@Q8_0
llama.cpp: llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 -c 8192
vLLM: vllm serve Qwen/Qwen3-1.7B --max-model-len 8192 --gpu-memory-utilization 0.9

Open Qwen3 1.7B in the checker

Hardware that fits

Hardware	Runtime	Quant	Memory	Fit	Speed	Command
M1 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~25 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 3060 Ti 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~122 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 3070 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~122 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4060 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~74 tok/s	`ollama run qwen3:1.7b-q8_0`
RX 7600 8GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~79 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 3080 10GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~207 tok/s	`ollama run qwen3:1.7b-q8_0`
M1 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~25 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~44 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 3060 12GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~98 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4070 12GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~137 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 5070 12GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~183 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4060 Ti 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~79 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4070 Ti Super 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~183 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4080 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~196 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 5080 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~262 tok/s	`ollama run qwen3:1.7b-q8_0`
RX 7800 XT 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~170 tok/s	`ollama run qwen3:1.7b-q8_0`
RX 9070 XT 16GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~175 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~36 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~44 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 3090 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~255 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 4090 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~275 tok/s	`ollama run qwen3:1.7b-q8_0`
RX 7900 XTX 24GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~262 tok/s	`ollama run qwen3:1.7b-q8_0`
M1 Pro 32GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~73 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 Pro 32GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~73 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 32GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~44 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 Pro 36GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~55 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M4 Max (32c GPU) 36GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~149 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Max (32c GPU) 36GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~167 tok/s	`ollama run qwen3:1.7b-q8_0`
RTX 5090 32GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~489 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 Pro 48GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~99 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M4 Max (40c GPU) 48GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~199 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Max (40c GPU) 48GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~223 tok/s	`ollama run qwen3:1.7b-q8_0`
M1 Max 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~145 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 Max 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~145 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 Max 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~145 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 Max 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~199 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M4 Max (40c GPU) 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~199 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Max (40c GPU) 64GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~223 tok/s	`ollama run qwen3:1.7b-q8_0`
DGX Spark 64GB (ASUS, Dell, HP, MSI, Acer, Gigabyte)	Ollama	Q8_0	2.9 GB	fits in fast memory	~93 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M3 Ultra 96GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~298 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Ultra 96GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~436 tok/s	`ollama run qwen3:1.7b-q8_0`
M3 Max 128GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~145 tok/s	`ollama run qwen3:1.7b-q8_0`
M4 Max 128GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~199 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M4 Max (40c GPU) 128GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~199 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Max (40c GPU) 128GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~223 tok/s	`ollama run qwen3:1.7b-q8_0`
DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte)	Ollama	Q8_0	2.9 GB	fits in fast memory	~93 tok/s	`ollama run qwen3:1.7b-q8_0`
M2 Ultra 192GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~291 tok/s	`ollama run qwen3:1.7b-q8_0`
2x Mac Studio M5 Max (40c GPU) 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~228 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node`
Mac Studio M3 Ultra 256GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~298 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Ultra (64c GPU) 256GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~436 tok/s	`ollama run qwen3:1.7b-q8_0`
Mac Studio M5 Ultra (80c GPU) 256GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~436 tok/s	`ollama run qwen3:1.7b-q8_0`
2x DGX Spark 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~95 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node`
4x Mac Studio M5 Max (40c GPU) 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~201 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node`
2x Mac Studio M5 Ultra (80c GPU) 256GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~446 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node`
Mac Studio M3 Ultra (80c GPU) 512GB	Ollama	Q8_0	2.9 GB	fits in fast memory	~298 tok/s	`ollama run qwen3:1.7b-q8_0`
4x DGX Spark 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~84 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node`
4x Mac Studio M5 Ultra (80c GPU) 256GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~393 tok/s	`llama-server -hf unsloth/Qwen3-1.7B-GGUF:Q8_0 --rpc node2:50052,node3:50052,node4:50052 -c 8192 # run rpc-server on each other node`

Estimates, not benchmarks. Model sizes are ollama.com download sizes; memory needed adds headroom for ~8k tokens of context; speeds are rough figures from memory bandwidth. Data updated 2026-10-16.

A project by OnDevice Labs · llms.txt