allenai.Llama 3.1 Tulu 3 405B needs about 247.8 GB of memory at Q4_K_M (a 243.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 5 quants, from 149.3 GB (Q2_K) to 332.9 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 16.9 GB.
allenai.Llama 3.1 Tulu 3 405B is a 406B-parameter Llama model in GGUF format, quantized by DevQuasar-7. Its trained context length is 128K tokens. The repository has 5 quant files, from 149.3 GB (Q2_K) to 332.9 GB (Q6_K).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 149.3 GB (11 parts) | 154.1 GB | Too big | Get file → | |
| Q3_K_M | 195.4 GB (15 parts) | 200.1 GB | Too big | Get file → | |
| Q4_K_M | 243.1 GB (18 parts) | 247.8 GB | Too big | Get file → | |
| Q5_K_M | 286.6 GB (22 parts) | 291.4 GB | Too big | Get file → | |
| Q6_K | 332.9 GB (25 parts) | 337.7 GB | Too big | Get file → |
What GPU do you need for allenai.Llama 3.1 Tulu 3 405B?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: no quant fits at an 8K context.
Which allenai.Llama 3.1 Tulu 3 405B quant should you choose?
Q4_K_M is the usual starting point for allenai.Llama 3.1 Tulu 3 405B: it balances file size and quality well. Here it is a 243.1 GB file.
If memory is tight, the smallest file is Q2_K at 149.3 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about allenai.Llama 3.1 Tulu 3 405B
How much VRAM does allenai.Llama 3.1 Tulu 3 405B need?
About 247.8 GB at Q4_K_M with an 8K context window. That is the 243.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run allenai.Llama 3.1 Tulu 3 405B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 154.1 GB, so you would need a larger GPU or CPU offloading.
What is the file size of allenai.Llama 3.1 Tulu 3 405B GGUF?
It depends on the quant: Q2_K is 149.3 GB, Q3_K_M is 195.4 GB, Q4_K_M is 243.1 GB, Q5_K_M is 286.6 GB, Q6_K is 332.9 GB.
What context length does allenai.Llama 3.1 Tulu 3 405B support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 4.2 GB at 8K and 16.9 GB at 32K.
Where can I download allenai.Llama 3.1 Tulu 3 405B GGUF?
The files are on Hugging Face in the DevQuasar-7/allenai.Llama-3.1-Tulu-3-405B-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does allenai.Llama 3.1 Tulu 3 405B use?
Hugging Face lists the license as llama3.1. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.