Models / Yi-1.5
Yi-1.5 by 01-ai
10 models (6B to 34B) · ↓ 111.8K downloads a month · from $0.19/h on 2x RTX 3060 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.
Yi-1.5 6B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-6B | vLLM | 12 GB | 16 GB | $0.19/h 2x RTX 3060 |
Yi-1.5 6B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-6B-Chat | vLLM | 12 GB | 16 GB | $0.19/h 2x RTX 3060 |
Yi-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-9B-Chat | vLLM | 18 GB | 22.3 GB | $0.20/h RTX 3090 |
Yi-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-9B | vLLM | 18 GB | 22.3 GB | $0.20/h RTX 3090 |
Yi-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-9B-Chat-16K | vLLM | 18 GB | 22.7 GB | $0.20/h RTX 3090 |
Yi-1.5 9B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-9B-32K | vLLM | 18 GB | 22.7 GB | $0.20/h RTX 3090 |
Yi-1.5 34B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-34B-Chat | vLLM | 69 GB | 79.8 GB | $0.89/h 2x Nvidia A40 |
Yi-1.5 34B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-34B | vLLM | 69 GB | 79.8 GB | $0.89/h 2x Nvidia A40 |
Yi-1.5 34B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-34B-Chat-16K | vLLM | 69 GB | 80.9 GB | $0.89/h 2x Nvidia A40 |
Yi-1.5 34B
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|
| Original01-ai/Yi-1.5-34B-32K | vLLM | 69 GB | 80.9 GB | $0.89/h 2x Nvidia A40 |
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").