TareksTesting Alkahest V3.2 LLaMa 70B needs about 45.7 GB of memory at Q4_K_M (a 42.5 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 10 quants, from 30.9 GB (Q3_K_S) to 75.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 10.7 GB.
TareksTesting Alkahest V3.2 LLaMa 70B is a 70.6B-parameter Llama model in GGUF format, quantized by ArtusDev. Its trained context length is 128K tokens. The repository has 10 quant files, from 30.9 GB (Q3_K_S) to 75.0 GB (Q8_0).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_S | 30.9 GB | 34.1 GB | Too big | Get file → | |
| Q3_K_M | 34.3 GB | 37.5 GB | Too big | Get file → | |
| Q3_K_L | 37.1 GB | 40.3 GB | Too big | Get file → | |
| IQ4_XS | 37.9 GB | 41.1 GB | Too big | Get file → | |
| Q4_K_S | 40.3 GB | 43.5 GB | Too big | Get file → | |
| Q4_K_M | 42.5 GB | 45.7 GB | Too big | Get file → | |
| Q5_K_S | 48.7 GB (2 parts) | 51.8 GB | Too big | Get file → | |
| Q5_K_M | 49.9 GB (2 parts) | 53.1 GB | Too big | Get file → | |
| Q6_K | 57.9 GB (2 parts) | 61.1 GB | Too big | Get file → | |
| Q8_0 | 75.0 GB (2 parts) | 78.2 GB | Too big | Get file → |
What GPU do you need for TareksTesting Alkahest V3.2 LLaMa 70B?
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: the best quant that fits is Q4_K_M (42.5 GB file, about 45.7 GB in use).
Which TareksTesting Alkahest V3.2 LLaMa 70B quant should you choose?
Q4_K_M is the usual starting point for TareksTesting Alkahest V3.2 LLaMa 70B: it balances file size and quality well. Here it is a 42.5 GB file.
The Q8_0 file is 75.0 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q3_K_S at 30.9 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about TareksTesting Alkahest V3.2 LLaMa 70B
How much VRAM does TareksTesting Alkahest V3.2 LLaMa 70B need?
About 45.7 GB at Q4_K_M with an 8K context window. That is the 42.5 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 TareksTesting Alkahest V3.2 LLaMa 70B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 34.1 GB, so you would need a larger GPU or CPU offloading.
What is the file size of TareksTesting Alkahest V3.2 LLaMa 70B GGUF?
It depends on the quant: Q3_K_S is 30.9 GB, Q3_K_M is 34.3 GB, Q3_K_L is 37.1 GB, IQ4_XS is 37.9 GB, Q4_K_S is 40.3 GB, Q4_K_M is 42.5 GB and 4 more.
What context length does TareksTesting Alkahest V3.2 LLaMa 70B support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 2.7 GB at 8K and 10.7 GB at 32K.
Where can I download TareksTesting Alkahest V3.2 LLaMa 70B GGUF?
The files are on Hugging Face in the ArtusDev/TareksTesting_Alkahest-V3.2-LLaMa-70B-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 TareksTesting Alkahest V3.2 LLaMa 70B use?
Hugging Face lists the license as llama3. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.