← CanIRunLocal · All hardware · All models
IBM · 3.66B parameters · chat, reasoning, coding.
Ollama weight files: Q4_K_M 2.2 GB (granite4.2:3b); Q5_K_M 2.6 GB (granite4.2:3b-q5_K_M); Q8_0 3.9 GB (granite4.2:3b-q8_0).
Yes on 85 machines in this list. The smallest is M1 8GB: Q8_0, about 4.7 GB, roughly 14 tokens per second, via Ollama. ollama run granite4.2:3b-q8_0
ollama run granite4.2:3b-q8_0ollama run granite4.2:3b-q8_0ollama run granite4.2:3b-q8_0ollama run granite4.2:3b-q8_0lms get ibm-granite/granite-4.2-3b-GGUF@Q8_0llama-server -hf ibm-granite/granite-4.2-3b-GGUF:Q8_0 -c 8192vllm serve ibm-granite/granite-4.2-3b --max-model-len 8192 --gpu-memory-utilization 0.9| Hardware | Runtime | Quant | Memory | Fit | Speed | Command |
|---|---|---|---|---|---|---|
| M1 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~14 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3060 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~37 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3060 Ti 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~69 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3070 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~69 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4060 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~42 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4060 Ti 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~44 tok/s | ollama run granite4.2:3b-q8_0 |
| RX 7600 8GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~44 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3080 10GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~117 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~14 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 Pro 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~41 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Pro 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~41 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~25 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3060 12GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~55 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3080 12GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~140 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4070 12GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~78 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 5070 12GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~103 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Pro 18GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~31 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4060 Ti 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~44 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4070 Ti Super 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~103 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4080 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~110 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 5080 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~148 tok/s | ollama run granite4.2:3b-q8_0 |
| RX 7800 XT 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~96 tok/s | ollama run granite4.2:3b-q8_0 |
| RX 9070 XT 16GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~98 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~21 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~25 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Pro 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~56 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 3090 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~144 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 4090 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~155 tok/s | ollama run granite4.2:3b-q8_0 |
| RX 7900 XTX 24GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~148 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 Max (24c GPU) 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 Max (32c GPU) 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 Pro 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~41 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Max (30c GPU) 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Max (38c GPU) 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Pro 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~41 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~25 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Max (30c GPU) 36GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~62 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Pro 36GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~31 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Max (32c GPU) 36GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~84 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M4 Max (32c GPU) 36GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~84 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Max (32c GPU) 36GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~94 tok/s | ollama run granite4.2:3b-q8_0 |
| RTX 5090 32GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~276 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Max (40c GPU) 48GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Max (40c GPU) 48GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Pro 48GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~56 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M4 Max (40c GPU) 48GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Max (40c GPU) 48GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~126 tok/s | ollama run granite4.2:3b-q8_0 |
| M1 Max (32c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Max (30c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Max (38c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Ultra 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~164 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Max (40c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Max (40c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Pro 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~56 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M4 Max (40c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Max (40c GPU) 64GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~126 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Max (38c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Max (30c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~62 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M3 Ultra (60c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~168 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M3 Ultra (80c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~168 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Ultra (64c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~246 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Ultra (80c GPU) 96GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~246 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Ultra 128GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~164 tok/s | ollama run granite4.2:3b-q8_0 |
| M3 Max (40c GPU) 128GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~82 tok/s | ollama run granite4.2:3b-q8_0 |
| M4 Max (40c GPU) 128GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M4 Max (40c GPU) 128GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~112 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Max (40c GPU) 128GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~126 tok/s | ollama run granite4.2:3b-q8_0 |
| DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte) | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~53 tok/s | ollama run granite4.2:3b-q8_0 |
| M2 Ultra 192GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~164 tok/s | ollama run granite4.2:3b-q8_0 |
| 2x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q8_0 GGUF | 4.7 GB | fits in fast memory | ~107 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-GGUF:Q8_0 --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra (60c GPU) 256GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~168 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M3 Ultra (80c GPU) 256GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~168 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Ultra (64c GPU) 256GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~246 tok/s | ollama run granite4.2:3b-q8_0 |
| Mac Studio M5 Ultra (80c GPU) 256GB | Ollama | Q8_0 | 4.7 GB | fits in fast memory | ~246 tok/s | ollama run granite4.2:3b-q8_0 |
| 2x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 4.7 GB | fits in fast memory | ~45 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-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 | 4.7 GB | fits in fast memory | ~95 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-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 | 4.7 GB | fits in fast memory | ~210 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-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 | 4.7 GB | fits in fast memory | ~168 tok/s | ollama run granite4.2:3b-q8_0 |
| 4x DGX Spark 128GB | llama.cpp | Q8_0 GGUF | 4.7 GB | fits in fast memory | ~39 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-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 | 4.7 GB | fits in fast memory | ~185 tok/s | llama-server -hf ibm-granite/granite-4.2-3b-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