← CanIRunLocal · All hardware · All models

Can I run SmolLM2 1.7B locally?

HuggingFace · 1.71B parameters · chat.

Ollama weight files: Q5_K_M 1.2 GB (smollm2:1.7b-instruct-q5_K_M); Q8_0 1.8 GB (smollm2:1.7b); Q4_K_M 1.1 GB (smollm2:1.7b-instruct-q4_K_M).

Yes on 62 machines in this list. The smallest is M1 8GB: Q8_0, about 2.5 GB, roughly 30 tokens per second, via Ollama. ollama run smollm2:1.7b

DGX Spark 128GB: Q8_0, ~2.5 GB, fits in fast memory, ~114 tok/s — ollama run smollm2:1.7b
Mac Studio M5 Max 128GB: Q8_0, ~2.5 GB, fits in fast memory, ~273 tok/s — ollama run smollm2:1.7b
Mac Studio M5 Ultra 256GB: Q8_0, ~2.5 GB, fits in fast memory, ~533 tok/s — ollama run smollm2:1.7b

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

Ollama: ollama run smollm2:1.7b
LM Studio: lms get unsloth/SmolLM2-1.7B-Instruct-GGUF@Q8_0
llama.cpp: llama-server -hf unsloth/SmolLM2-1.7B-Instruct-GGUF:Q8_0 -c 8192
vLLM: vllm serve HuggingFaceTB/SmolLM2-1.7B-Instruct --max-model-len 8192 --gpu-memory-utilization 0.9

Open SmolLM2 1.7B in the checker

Hardware that fits

Hardware	Runtime	Quant	Memory	Fit	Speed	Command
M1 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~30 tok/s	`ollama run smollm2:1.7b`
M2 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
M3 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
RTX 3060 Ti 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~149 tok/s	`ollama run smollm2:1.7b`
RTX 3070 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~149 tok/s	`ollama run smollm2:1.7b`
RTX 4060 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~91 tok/s	`ollama run smollm2:1.7b`
RX 7600 8GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~96 tok/s	`ollama run smollm2:1.7b`
RTX 3080 10GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~253 tok/s	`ollama run smollm2:1.7b`
M1 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~30 tok/s	`ollama run smollm2:1.7b`
M2 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
M3 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
M4 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~53 tok/s	`ollama run smollm2:1.7b`
RTX 3060 12GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~120 tok/s	`ollama run smollm2:1.7b`
RTX 4070 12GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~168 tok/s	`ollama run smollm2:1.7b`
RTX 5070 12GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~224 tok/s	`ollama run smollm2:1.7b`
RTX 4060 Ti 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~96 tok/s	`ollama run smollm2:1.7b`
RTX 4070 Ti Super 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~224 tok/s	`ollama run smollm2:1.7b`
RTX 4080 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~239 tok/s	`ollama run smollm2:1.7b`
RTX 5080 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~320 tok/s	`ollama run smollm2:1.7b`
RX 7800 XT 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~208 tok/s	`ollama run smollm2:1.7b`
RX 9070 XT 16GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~213 tok/s	`ollama run smollm2:1.7b`
M2 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
M3 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~44 tok/s	`ollama run smollm2:1.7b`
M4 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~53 tok/s	`ollama run smollm2:1.7b`
RTX 3090 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~312 tok/s	`ollama run smollm2:1.7b`
RTX 4090 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~336 tok/s	`ollama run smollm2:1.7b`
RX 7900 XTX 24GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~320 tok/s	`ollama run smollm2:1.7b`
M1 Pro 32GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~89 tok/s	`ollama run smollm2:1.7b`
M2 Pro 32GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~89 tok/s	`ollama run smollm2:1.7b`
M4 32GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~53 tok/s	`ollama run smollm2:1.7b`
M3 Pro 36GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~67 tok/s	`ollama run smollm2:1.7b`
Mac Studio M4 Max (32c GPU) 36GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~182 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Max (32c GPU) 36GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~204 tok/s	`ollama run smollm2:1.7b`
RTX 5090 32GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~597 tok/s	`ollama run smollm2:1.7b`
M4 Pro 48GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~121 tok/s	`ollama run smollm2:1.7b`
Mac Studio M4 Max (40c GPU) 48GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~243 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Max (40c GPU) 48GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~273 tok/s	`ollama run smollm2:1.7b`
M1 Max 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~178 tok/s	`ollama run smollm2:1.7b`
M2 Max 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~178 tok/s	`ollama run smollm2:1.7b`
M3 Max 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~178 tok/s	`ollama run smollm2:1.7b`
M4 Max 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~243 tok/s	`ollama run smollm2:1.7b`
Mac Studio M4 Max (40c GPU) 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~243 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Max (40c GPU) 64GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~273 tok/s	`ollama run smollm2:1.7b`
DGX Spark 64GB (ASUS, Dell, HP, MSI, Acer, Gigabyte)	Ollama	Q8_0	2.5 GB	fits in fast memory	~114 tok/s	`ollama run smollm2:1.7b`
Mac Studio M3 Ultra 96GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~364 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Ultra 96GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~533 tok/s	`ollama run smollm2:1.7b`
M3 Max 128GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~178 tok/s	`ollama run smollm2:1.7b`
M4 Max 128GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~243 tok/s	`ollama run smollm2:1.7b`
Mac Studio M4 Max (40c GPU) 128GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~243 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Max (40c GPU) 128GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~273 tok/s	`ollama run smollm2:1.7b`
DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte)	Ollama	Q8_0	2.5 GB	fits in fast memory	~114 tok/s	`ollama run smollm2:1.7b`
M2 Ultra 192GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~356 tok/s	`ollama run smollm2:1.7b`
2x Mac Studio M5 Max (40c GPU) 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~229 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node`
Mac Studio M3 Ultra 256GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~364 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Ultra (64c GPU) 256GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~533 tok/s	`ollama run smollm2:1.7b`
Mac Studio M5 Ultra (80c GPU) 256GB	Ollama	Q8_0	2.5 GB	fits in fast memory	~533 tok/s	`ollama run smollm2:1.7b`
2x DGX Spark 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~96 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-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	~202 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-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	~448 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-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	2.5 GB	fits in fast memory	~364 tok/s	`ollama run smollm2:1.7b`
4x DGX Spark 128GB	llama.cpp	Q8_0 GGUF	2.5 GB	fits in fast memory	~84 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-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	~396 tok/s	`llama-server -hf unsloth/SmolLM2-1.7B-Instruct-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 2026-10-18.

A project by OnDevice Labs · llms.txt