Models / Ornith-1.5
Ornith-1.5 by ornith-ai
8 models (9B to 397B) · ↓ 4.3M downloads a month · from $0.14/h on RTX A4000 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
Ornith-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-9B | vLLM | 19 GB | 24.9 GB | $0.34/h 2x RTX 5060 Ti |
| OBLITERATED (OBLITERATUS)OBLITERATUS/Ornith-1.5-9B-OBLITERATED | vLLM | 19 GB | 24.9 GB | $0.34/h 2x RTX 5060 Ti |
Ornith-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-9B-NVFP4 | vLLM | 8.8 GB | 13.2 GB | $0.14/h RTX A4000 |
Ornith-1.5 35B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-35B-A3B-NVFP4 | vLLM | 23 GB | 29 GB | $0.34/h 2x RTX 5060 Ti |
Ornith-1.5 35B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-35B-A3B-FP8 | vLLM | 39 GB | 46.7 GB | $0.46/h RTX A6000 |
Ornith-1.5 35B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-35B-A3B | vLLM | 72 GB | 82.9 GB | $0.89/h 2x Nvidia A40 |
| AWQ INT4 (cyankiwi)cyankiwi/Ornith-1.5-35B-A3B-AWQ-INT4 | vLLM | 26 GB | 32 GB | $0.45/h Nvidia A40 |
Ornith-1.5 397B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-397B-NVFP4 | vLLM | 242 GB | 272 GB | $5.33/h 8x Nvidia A100 |
Ornith-1.5 397B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-397B-FP8 | vLLM | 418 GB | 468.1 GB | $11.77/h 8x RTX PRO 6000 |
Ornith-1.5 397B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Originalornith-ai/Ornith-1.5-397B | vLLM | 807 GB | 899.8 GB | $44.03/h 8x Nvidia H200 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").