OS Shepherd 9B needs about 7.5 GB of memory at Q4_K_M (a 5.9 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 25 quants, from 0.0 GB to 17.9 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.
OS Shepherd 9B – multimodal reward model for evaluating agent trajectories
OS Shepherd 9B is a 9B parameter multimodal reward model fine‑tuned from Qwen3.5‑9B. It judges computer‑use agent trajectories, returning a reasoned SUCCESS or FAIL verdict. The model supports text and image input and is licensed under Apache‑2.0.
Good at
- Trajectory evaluation
- Reward‑model research
- Data filtering
Limits
- No vision support beyond image input
- Requires recent llama.cpp release b10262
- Large context length 262144 tokens
Best for
- Evaluating agent task completion
- Filtering trajectory datasets
- Researching reward models
Who it’s for. Developers and researchers working on computer‑use agents who need a reliable reward model for trajectory evaluation. The model is not a control policy and may miss fine‑grained visual failures.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 1.6 GB | Fits | Get file → | |
| IQ2_M | 3.8 GB | 5.3 GB | Fits | Get file → | |
| Q2_K | 4.1 GB | 5.6 GB | Fits | Get file → | |
| IQ3_XXS | 4.3 GB | 5.8 GB | Fits | Get file → | |
| IQ3_XS | 4.6 GB | 6.1 GB | Fits | Get file → | |
| Q3_K_S | 4.7 GB | 6.2 GB | Fits | Get file → | |
| IQ3_M | 4.7 GB | 6.3 GB | Fits | Get file → | |
| Q3_K_M | 4.9 GB | 6.5 GB | Fits | Get file → | |
| Q2_K_L | 5.1 GB | 6.6 GB | Fits | Get file → | |
| Q3_K_L | 5.1 GB | 6.7 GB | Fits | Get file → | |
| IQ4_XS | 5.2 GB | 6.8 GB | Fits | Get file → | |
| IQ4_NL | 5.5 GB | 7.1 GB | Fits | Get file → | |
| Q4_0 | 5.5 GB | 7.1 GB | Fits | Get file → | |
| Q4_K_S | 5.6 GB | 7.2 GB | Fits | Get file → | |
| Q4_K_M | 5.9 GB | 7.5 GB | Tight | Get file → | |
| Q4_1 | 5.9 GB | 7.5 GB | Tight | Get file → | |
| Q3_K_XL | 6.0 GB | 7.6 GB | Tight | Get file → | |
| Q5_K_S | 6.5 GB | 8.1 GB | Too big | Get file → | |
| Q4_K_L | 6.7 GB | 8.2 GB | Too big | Get file → | |
| Q5_K_M | 6.9 GB | 8.4 GB | Too big | Get file → | |
| Q5_K_L | 7.5 GB | 9.1 GB | Too big | Get file → | |
| Q6_K | 7.7 GB | 9.3 GB | Too big | Get file → | |
| Q6_K_L | 8.2 GB | 9.8 GB | Too big | Get file → | |
| Q8_0 | 9.5 GB | 11.1 GB | Too big | Get file → | |
| BF16 | 17.9 GB | 19.5 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (5.91 GB) or Q4_K_S (5.6 GB). For 12 GB VRAM, Q5_K_S (6.53 GB) or Q5_K_M (6.85 GB) provide a good balance. Larger models such as Q8_0 (9.55 GB) need 16 GB or more.
What GPU do you need for OS Shepherd 9B?
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: the best quant that fits is the main file (0.0 GB file, about 1.6 GB in use).
- 6 GB VRAM: the best quant that fits is IQ3_XXS (4.3 GB file, about 5.8 GB in use).
- 8 GB VRAM: the best quant that fits is Q3_K_XL (6.0 GB file, about 7.6 GB in use).
- 12 GB VRAM: the best quant that fits is Q8_0 (9.5 GB file, about 11.1 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (9.5 GB file, about 11.1 GB in use).
- 24 GB VRAM: the best quant that fits is BF16 (17.9 GB file, about 19.5 GB in use).
Which OS Shepherd 9B quant should you choose?
Q4_K_M is the usual starting point for OS Shepherd 9B: it balances file size and quality well. Here it is a 5.9 GB file.
The Q8_0 file is 9.5 GB, about 1.6 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 the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about OS Shepherd 9B
How much VRAM does OS Shepherd 9B need?
About 7.5 GB at Q4_K_M with an 8K context window. That is the 5.9 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 OS Shepherd 9B on an 8 GB GPU?
Yes, with Q3_K_XL: the 6.0 GB file needs about 7.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of OS Shepherd 9B GGUF?
It depends on the quant: file is 0.0 GB, IQ2_M is 3.8 GB, Q2_K is 4.1 GB, IQ3_XXS is 4.3 GB, IQ3_XS is 4.6 GB, Q3_K_S is 4.7 GB and 19 more.
What context length does OS Shepherd 9B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.
Where can I download OS Shepherd 9B GGUF?
The files are on Hugging Face in the bartowski/OS-Copilot_OS-Shepherd-9B-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 OS Shepherd 9B use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run OS Shepherd 9B on 8 GB VRAM?
Yes, use Q4_K_M (5.91 GB) or Q4_K_S (5.6 GB) quantizations. They fit within 8 GB VRAM with some headroom.
Does the model support image input?
Yes, it accepts image input via the mmproj projector files that pair with any quant.
What license applies to this model?
The model is released under Apache License 2.0.