No — not on this device
Ornith 1.5 397B-A17B at Q4_K_M needs 224.1 GB against 204.8 GB usable, and the shortfall of 19.3 GB is more than 16 GB of system RAM can cover at a tolerable speed. A smaller sibling or a lower quantisation is the honest answer here.
The flagship Ornith on the Qwen3.5 397B-A17B architecture, MIT-licensed. A 256 GB Mac Studio at Q4, and it is in the Ollama library.
What hardware do I need for Ornith 1.5 397B-A17B? →
Fits instead: Ornith 1.5 35B-A3B (20.9 GB) · Ornith 1.5 9B (6.1 GB)
The VRAM budget
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 392.8 GB | 393.7 GB | — | ~3.4 | −0.1% ppl | 188.9 GB over |
| Q6_K | 303.2 GB | 304.0 GB | — | ~4.4 | −0.4% ppl | 99.2 GB over |
| Q5_K_M | 262.0 GB | 262.9 GB | — | ~5.1 | −0.8% ppl | 58.1 GB over |
| Q4_K_M | 223.2 GB | 224.1 GB | — | ~6.0 | −1.9% ppl | 19.3 GB over |
| Q3_K_M | 180.7 GB | 181.5 GB | 256K | 7.4 | −5.4% ppl | Long context |
| Q2_K | 154.8 GB | 155.7 GB | 256K | 8.6 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 15 of its 60 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state.
How to run it
$ llama-server \
-hf ornith-ai/Ornith-1.5-397B:Q4_K_M \
-c 8192 -ngl 54
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.