Ornith-1.0-397B: hardware requirements
Ornith-1.0-397B needs about 246 GB of memory at Q4 quantization — a 242 GB download plus room for context and overhead. That is beyond consumer hardware; use it through an API instead.
This is a mixture-of-experts model. Only a fraction of its 397B parameters are used for each token, so it runs faster than its size suggests — but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →
Will it run on your machine?
Download sizes by quantization
Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?
| Quant | Download | RAM @ 4K | RAM @ 32K | Source |
|---|---|---|---|---|
| Q4_K_Mrecommended | 242 GB | 246 GB | 249 GB | measured |
| Q5_K_M | 283 GB | 287 GB | 291 GB | measured |
| Q6_K | 342 GB | 346 GB | 350 GB | measured |
| Q8_0 | 422 GB | 425 GB | 429 GB | measured |
| BF16 | 794 GB | 797 GB | 801 GB | estimated |
“Measured” sizes are read from published GGUF files (bartowski/deepreinforce-ai_Ornith-1.0-397B-GGUF). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.