Tunguska 39B V1A Workspace needs about 27.2 GB of memory at Q4_K_M (a 23.4 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 25 quants, from 10.4 GB (IQ2_XXS) to 78.0 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 13.3 GB.
Tunguska 39B V1A Workspace is a 39B-parameter Tunguska model in GGUF format, quantized by bartowski. Its trained context length is 128K tokens. The repository has 25 quant files, from 10.4 GB (IQ2_XXS) to 78.0 GB (F16).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_XXS | 10.4 GB | 14.3 GB | Too big | Get file → | |
| IQ2_XS | 11.6 GB | 15.4 GB | Too big | Get file → | |
| IQ2_S | 12.3 GB | 16.1 GB | Too big | Get file → | |
| IQ2_M | 13.3 GB | 17.1 GB | Too big | Get file → | |
| Q2_K | 14.4 GB | 18.3 GB | Too big | Get file → | |
| Q2_K_L | 14.6 GB | 18.5 GB | Too big | Get file → | |
| IQ3_XS | 16.0 GB | 19.9 GB | Too big | Get file → | |
| Q3_K_S | 16.9 GB | 20.7 GB | Too big | Get file → | |
| IQ3_M | 17.6 GB | 21.4 GB | Too big | Get file → | |
| Q3_K_M | 18.8 GB | 22.7 GB | Too big | Get file → | |
| Q3_K_L | 20.5 GB | 24.4 GB | Too big | Get file → | |
| Q3_K_XL | 20.7 GB | 24.5 GB | Too big | Get file → | |
| IQ4_XS | 20.9 GB | 24.7 GB | Too big | Get file → | |
| Q4_0 | 22.1 GB | 25.9 GB | Too big | Get file → | |
| IQ4_NL | 22.1 GB | 25.9 GB | Too big | Get file → | |
| Q4_K_S | 22.2 GB | 26.0 GB | Too big | Get file → | |
| Q4_K_M | 23.4 GB | 27.2 GB | Too big | Get file → | |
| Q4_K_L | 23.5 GB | 27.3 GB | Too big | Get file → | |
| Q5_K_S | 26.9 GB | 30.7 GB | Too big | Get file → | |
| Q5_K_M | 27.6 GB | 31.4 GB | Too big | Get file → | |
| Q5_K_L | 27.7 GB | 31.5 GB | Too big | Get file → | |
| Q6_K | 32.0 GB | 35.8 GB | Too big | Get file → | |
| Q6_K_L | 32.1 GB | 35.9 GB | Too big | Get file → | |
| Q8_0 | 41.5 GB | 45.3 GB | Too big | Get file → | |
| F16 | 78.0 GB (2 parts) | 81.9 GB | Too big | Get file → |
What GPU do you need for Tunguska 39B V1A Workspace?
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: the best quant that fits is IQ2_XS (11.6 GB file, about 15.4 GB in use).
- 24 GB VRAM: the best quant that fits is Q3_K_M (18.8 GB file, about 22.7 GB in use).
- 32 GB VRAM: the best quant that fits is Q5_K_L (27.7 GB file, about 31.5 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (41.5 GB file, about 45.3 GB in use).
Which Tunguska 39B V1A Workspace quant should you choose?
Q4_K_M is the usual starting point for Tunguska 39B V1A Workspace: it balances file size and quality well. Here it is a 23.4 GB file.
The Q8_0 file is 41.5 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 IQ2_XXS at 10.4 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 Tunguska 39B V1A Workspace
How much VRAM does Tunguska 39B V1A Workspace need?
About 27.2 GB at Q4_K_M with an 8K context window. That is the 23.4 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 Tunguska 39B V1A Workspace on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 14.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Tunguska 39B V1A Workspace GGUF?
It depends on the quant: IQ2_XXS is 10.4 GB, IQ2_XS is 11.6 GB, IQ2_S is 12.3 GB, IQ2_M is 13.3 GB, Q2_K is 14.4 GB, Q2_K_L is 14.6 GB and 19 more.
What context length does Tunguska 39B V1A Workspace support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 3.3 GB at 8K and 13.3 GB at 32K.
Where can I download Tunguska 39B V1A Workspace GGUF?
The files are on Hugging Face in the bartowski/Tunguska-39B-v1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.