Can Ornith 1.5 35B A3B run on the RTX 6000 Ada?
Yes. At 8K context the best fit is Q8_0, using about 36.2 GB of the 45.6 GB usable on the RTX 6000 Ada. Estimated speed ceiling: ~304 tokens/sec.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 13.1 GB Fits | 13.2 GB Fits | 13.7 GB Fits | 15.6 GB Fits | 879 |
| 3-bit (Q3_K_M) • | 16.5 GB Fits | 16.6 GB Fits | 17.0 GB Fits | 18.9 GB Fits | 689 |
| 4-bit (Q4_K_M) • | 21.3 GB Fits | 21.4 GB Fits | 21.9 GB Fits | 23.7 GB Fits | 526 |
| 5-bit (Q5_K_M) • | 24.7 GB Fits | 24.8 GB Fits | 25.2 GB Fits | 27.1 GB Fits | 451 |
| 6-bit (Q6_K) • | 29.4 GB Fits | 29.5 GB Fits | 29.9 GB Fits | 31.8 GB Fits | 377 |
| 8-bit (Q8_0) • | 36.2 GB Fits | 36.2 GB Fits | 36.7 GB Fits | 38.6 GB Fits | 304 |
| 16-bit (F16/BF16) • | 67.1 GB No | 67.2 GB No | 67.7 GB No | 69.6 GB No | — |
• real GGUF file size from bartowski/Ornith-1.5-35B-A3B-GGUF. Speed is a bandwidth ceiling (960 GB/s), not a benchmark. Methodology.
More
Ornith 1.5 35B A3B on every GPU and Mac · Every model that runs on the RTX 6000 Ada · Other models that fit 48 GB VRAM · Try other settings in the calculator
Data: Hugging Face Hub (ornith-ai/Ornith-1.5-35B-A3B), NVIDIA specs. Data as of 2026-11-17.