← CanIRunLocal · All hardware · All models
NVIDIA · 3.98B parameters · chat.
Ollama weight files: Q5_K_M 3.67 GB (hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M, size from the Hugging Face file listing).
Yes on 85 machines in this list. The smallest is M1 8GB: Q5_K_M, about 4.5 GB, roughly 15 tokens per second, via Ollama. ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M
ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_Mollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_Mollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_Mollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_Mlms get nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF@Q5_K_Mllama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M -c 8192vllm serve nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 --max-model-len 8192 --gpu-memory-utilization 0.9| Hardware | Runtime | Quant | Memory | Fit | Speed | Command |
|---|---|---|---|---|---|---|
| M1 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~15 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3060 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~39 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3060 Ti 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~73 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3070 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~73 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4060 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~44 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4060 Ti 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~47 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RX 7600 8GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~47 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3080 10GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~124 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~15 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 Pro 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~44 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Pro 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~44 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~26 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3060 12GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~59 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3080 12GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~149 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4070 12GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~82 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 5070 12GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~110 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Pro 18GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~33 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4060 Ti 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~47 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4070 Ti Super 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~110 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4080 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~117 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 5080 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~157 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RX 7800 XT 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~102 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RX 9070 XT 16GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~105 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~22 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~26 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Pro 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~60 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 3090 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~153 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 4090 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~165 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RX 7900 XTX 24GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~157 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 Max (24c GPU) 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 Max (32c GPU) 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 Pro 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~44 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Max (30c GPU) 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Max (38c GPU) 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Pro 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~44 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~26 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Max (30c GPU) 36GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~65 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Pro 36GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~33 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Max (32c GPU) 36GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~89 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M4 Max (32c GPU) 36GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~89 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Max (32c GPU) 36GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~100 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| RTX 5090 32GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~293 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Max (40c GPU) 48GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Max (40c GPU) 48GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Pro 48GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~60 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M4 Max (40c GPU) 48GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Max (40c GPU) 48GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~134 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M1 Max (32c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Max (30c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Max (38c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Ultra 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~174 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Max (40c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Max (40c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Pro 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~60 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M4 Max (40c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Max (40c GPU) 64GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~134 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Max (38c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Max (30c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~65 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M3 Ultra (60c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~179 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M3 Ultra (80c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~179 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Ultra (64c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~262 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Ultra (80c GPU) 96GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~262 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Ultra 128GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~174 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M3 Max (40c GPU) 128GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~87 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M4 Max (40c GPU) 128GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M4 Max (40c GPU) 128GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~119 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Max (40c GPU) 128GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~134 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| DGX Spark 128GB (NVIDIA, ASUS, Dell, Lenovo, HP, MSI, Acer, Gigabyte) | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~56 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| M2 Ultra 192GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~174 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| 2x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~114 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra (60c GPU) 256GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~179 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M3 Ultra (80c GPU) 256GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~179 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Ultra (64c GPU) 256GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~262 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| Mac Studio M5 Ultra (80c GPU) 256GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~262 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| 2x DGX Spark 128GB | llama.cpp | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~47 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| 4x Mac Studio M5 Max (40c GPU) 128GB | llama.cpp | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~100 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --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 | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~222 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --rpc node2:50052 -c 8192 # run rpc-server on each other node |
| Mac Studio M3 Ultra (80c GPU) 512GB | Ollama | Q5_K_M | 4.5 GB | fits in fast memory | ~179 tok/s | ollama run hf.co/nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M |
| 4x DGX Spark 128GB | llama.cpp | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~42 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --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 | Q5_K_M GGUF | 4.5 GB | fits in fast memory | ~196 tok/s | llama-server -hf nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUF:Q5_K_M --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 2026-10-20.
A project by OnDevice Labs · llms.txt