← CanIRunLocal · All hardware · All models
HuggingFace · 0.362B parameters · chat.
Ollama weight files: Q5_K_M 0.29 GB (smollm2:360m-instruct-q5_K_M); Q8_0 0.39 GB (smollm2:360m-instruct-q8_0); FP16 0.73 GB (smollm2:360m); Q4_K_M 0.27 GB (smollm2:360m-instruct-q4_K_M).
Yes on 62 machines in this list. The smallest is M1 8GB: FP16, about 1.3 GB, roughly 75 tokens per second, via Ollama. ollama run smollm2:360m
ollama run smollm2:360mollama run smollm2:360mollama run smollm2:360mollama run smollm2:360mlms get unsloth/SmolLM2-360M-Instruct-GGUF@Q8_0llama-server -hf unsloth/SmolLM2-360M-Instruct-GGUF:Q8_0 -c 8192vllm serve HuggingFaceTB/SmolLM2-360M-Instruct --max-model-len 8192 --gpu-memory-utilization 0.9| Hardware | Runtime | Quant | Memory | Fit | Speed | Command |
|---|---|---|---|---|---|---|
| M1 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~75 tok/s | ollama run smollm2:360m |
| M2 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| M3 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| RTX 3060 Ti 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~368 tok/s | ollama run smollm2:360m |
| RTX 3070 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~368 tok/s | ollama run smollm2:360m |
| RTX 4060 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~224 tok/s | ollama run smollm2:360m |
| RX 7600 8GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~237 tok/s | ollama run smollm2:360m |
| RTX 3080 10GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~625 tok/s | ollama run smollm2:360m |
| M1 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~75 tok/s | ollama run smollm2:360m |
| M2 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| M3 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| M4 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~132 tok/s | ollama run smollm2:360m |
| RTX 3060 12GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~296 tok/s | ollama run smollm2:360m |
| RTX 4070 12GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~414 tok/s | ollama run smollm2:360m |
| RTX 5070 12GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~552 tok/s | ollama run smollm2:360m |
| RTX 4060 Ti 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~237 tok/s | ollama run smollm2:360m |
| RTX 4070 Ti Super 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~552 tok/s | ollama run smollm2:360m |
| RTX 4080 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~589 tok/s | ollama run smollm2:360m |
| RTX 5080 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~789 tok/s | ollama run smollm2:360m |
| RX 7800 XT 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~513 tok/s | ollama run smollm2:360m |
| RX 9070 XT 16GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~526 tok/s | ollama run smollm2:360m |
| M2 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| M3 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~110 tok/s | ollama run smollm2:360m |
| M4 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~132 tok/s | ollama run smollm2:360m |
| RTX 3090 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~769 tok/s | ollama run smollm2:360m |
| RTX 4090 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~828 tok/s | ollama run smollm2:360m |
| RX 7900 XTX 24GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~789 tok/s | ollama run smollm2:360m |
| M1 Pro 32GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~219 tok/s | ollama run smollm2:360m |
| M2 Pro 32GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~219 tok/s | ollama run smollm2:360m |
| M4 32GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~132 tok/s | ollama run smollm2:360m |
| M3 Pro 36GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~164 tok/s | ollama run smollm2:360m |
| Mac Studio M4 Max (32c GPU) 36GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~449 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Max (32c GPU) 36GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~504 tok/s | ollama run smollm2:360m |
| RTX 5090 32GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~1473 tok/s | ollama run smollm2:360m |
| M4 Pro 48GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~299 tok/s | ollama run smollm2:360m |
| Mac Studio M4 Max (40c GPU) 48GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~598 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Max (40c GPU) 48GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~673 tok/s | ollama run smollm2:360m |
| M1 Max 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~438 tok/s | ollama run smollm2:360m |
| M2 Max 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~438 tok/s | ollama run smollm2:360m |
| M3 Max 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~438 tok/s | ollama run smollm2:360m |
| M4 Max 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~598 tok/s | ollama run smollm2:360m |
| Mac Studio M4 Max (40c GPU) 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~598 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Max (40c GPU) 64GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~673 tok/s | ollama run smollm2:360m |
| DGX Spark 64GB (ASUS, Dell, HP, MSI, Acer, Gigabyte) | Ollama | FP16 | 1.3 GB | fits in fast memory | ~280 tok/s | ollama run smollm2:360m |
| Mac Studio M3 Ultra 96GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~898 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Ultra 96GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~1315 tok/s | ollama run smollm2:360m |
| M3 Max 128GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~438 tok/s | ollama run smollm2:360m |
| M4 Max 128GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~598 tok/s | ollama run smollm2:360m |
| Mac Studio M4 Max (40c GPU) 128GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~598 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Max (40c GPU) 128GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~673 tok/s | ollama run smollm2:360m |
| DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte) | Ollama | FP16 | 1.3 GB | fits in fast memory | ~280 tok/s | ollama run smollm2:360m |
| M2 Ultra 192GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~877 tok/s | ollama run smollm2:360m |
| 2x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q8_0 GGUF | 1 GB | fits in fast memory | ~1071 tok/s | llama-server -hf unsloth/SmolLM2-360M-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra 256GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~898 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Ultra (64c GPU) 256GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~1315 tok/s | ollama run smollm2:360m |
| Mac Studio M5 Ultra (80c GPU) 256GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~1315 tok/s | ollama run smollm2:360m |
| 2x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 1 GB | fits in fast memory | ~446 tok/s | llama-server -hf unsloth/SmolLM2-360M-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 | 1 GB | fits in fast memory | ~945 tok/s | llama-server -hf unsloth/SmolLM2-360M-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 | 1 GB | fits in fast memory | ~2092 tok/s | llama-server -hf unsloth/SmolLM2-360M-Instruct-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra (80c GPU) 512GB | Ollama | FP16 | 1.3 GB | fits in fast memory | ~898 tok/s | ollama run smollm2:360m |
| 4x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 1 GB | fits in fast memory | ~394 tok/s | llama-server -hf unsloth/SmolLM2-360M-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 | 1 GB | fits in fast memory | ~1846 tok/s | llama-server -hf unsloth/SmolLM2-360M-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