Model family · 4 sizes
Ornith: which size runs locally?
Ornith comes in 4 sizes, from 9B to 35B. Some sizes are Mixture-of-Experts, so they run faster than their memory footprint suggests. Here is each size with its Q4_K_M weight, the memory it needs, and the hardware that runs it.
- Sizes
- 4
- Smallest
- 9B
- Largest
- 35B
- Runs from
- 8GB
The Ornith lineup
"Needs" is the sourced minimum memory for Q4_K_M with a small context. Larger context needs more.
Which Ornith fits your memory
Largest that fits: Ornith 1.5 9B (9B), best case on Nvidia GeForce RTX 3060 Ti (8GB).
Largest that fits: Ornith 1.5 9B (9B), best case on Nvidia GeForce RTX 4080 (16GB).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Nvidia GeForce RTX 4090 (24GB). Comfortable up to Ornith 1.5 9B (9B).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Nvidia GeForce RTX 5090 (32GB).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Apple M5 Pro (48GB).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Apple M4 Max (64GB).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Apple M5 Max (128GB).
Largest that fits: Ornith 1.5 35B-A3B (35B), best case on Apple M3 Ultra (256GB).
Best case means the most capable device at that size (usually a discrete GPU). A Mac at the same size sits roughly one rung lower; see the per-size breakdown on each memory budget page.
FAQ
Which Ornith size should I run locally?
Pick the largest size your memory allows. On 8GB (best case) up to Ornith 1.5 9B; On 16GB (best case) up to Ornith 1.5 9B; On 24GB (best case) up to Ornith 1.5 35B-A3B; On 32GB (best case) up to Ornith 1.5 35B-A3B; On 48GB (best case) up to Ornith 1.5 35B-A3B; On 64GB (best case) up to Ornith 1.5 35B-A3B; On 128GB (best case) up to Ornith 1.5 35B-A3B; On 256GB (best case) up to Ornith 1.5 35B-A3B. Smaller sizes run faster and leave headroom for context.
What is the smallest Ornith model?
Ornith 1.0 9B at 9B parameters, about 5.6 GB on disk at Q4_K_M. It is the one to use on phones and 8 GB machines.
What is the largest Ornith model and what does it need?
Ornith 1.5 35B-A3B at 35B (mixture of experts), about 20.22 GB at Q4_K_M. It fits a high-memory desktop GPU or Mac.
Understand the numbers
Short guides to the ideas behind Ornith's memory and quant figures.
Sources
- deep-reinforce.com
- huggingface.co/deepreinforce-ai/Ornith-1.0-35B
- huggingface.co/deepreinforce-ai/Ornith-1.0-35B-GGUF
- huggingface.co/deepreinforce-ai/Ornith-1.0-9B
- huggingface.co/deepreinforce-ai/Ornith-1.0-9B-GGUF
- huggingface.co/ornith-ai/Ornith-1.5-35B-A3B
- huggingface.co/ornith-ai/Ornith-1.5-35B-A3B-GGUF
- huggingface.co/ornith-ai/Ornith-1.5-9B
- huggingface.co/ornith-ai/Ornith-1.5-9B-GGUF
- ollama.com/library/ornith
- ollama.com/library/ornith-1.5
Memory figures are estimates at Q4_K_M. See methodology.