Can I run Ornith-1.5-122B on an RTX PRO 6000 Blackwell?
Yes, with a GGUF version
The published weights are too big, but the UD-Q4_K_XL file from unsloth/Ornith-1.5-122B-GGUF (79.7 GB) fits, with room for ≈ 256K tokens of context.
Full analysis of Ornith-1.5-122B
Everything that runs on RTX PRO 6000 Blackwell
Get the UD-Q4_K_XL file on Hugging Face
llama-server -hf unsloth/Ornith-1.5-122B-GGUF:UD-Q4_K_XLFiles and context on RTX PRO 6000 Blackwell
Each file of Ornith-1.5-122B (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.
| File | Size | On RTX PRO 6000 Blackwell | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| As published | 250 GB | Doesn't fit | — | — |
| BF16 | 249 GB | Doesn't fit | — | — |
| Q8_0 | 133 GB | Doesn't fit | — | — |
| UD-Q6_K_XL | 114 GB | Doesn't fit | — | — |
| UD-Q5_K_XL | 94.6 GB | Doesn't fit | — | — |
| UD-Q4_K_XL | 79.7 GB | Tight | 256K | 256K |
| UD-Q3_K_XL | 59.8 GB | Fits | 256K | 256K |
| UD-IQ3_XXS | 49.5 GB | Fits | 256K | 256K |
| UD-Q2_K_XL | 44.5 GB | Fits | 256K | 256K |
| UD-IQ2_XXS | 42.0 GB | Fits | 256K | 256K |
| UD-IQ1_M | 40.4 GB | Fits | 256K | 256K |
| BF16 | 249 GB | Doesn't fit | — | — |
| Q8_0 | 133 GB | Doesn't fit | — | — |
| Q6_K | 102 GB | Doesn't fit | — | — |
| Q5_K_M | 88.7 GB | Doesn't fit | — | — |
| Q4_K_M | 75.8 GB | Tight | 256K | 256K |