---
base_model:
- prithivMLmods/oMEGA-4B-SpatialThink-0804
tags:
- text-generation-inference
- spatial-reasoning
- llama-cpp
- vision-language
- multimodal
- image-captioning
- visual-question-answering
- conditional-generation
- vision
- language-model
- sft
- fine-grained-captioning
- computer-vision
datasets:
- prithivMLmods/OpenCaption-FineGrained
- remyxai/SpaceThinker
license: apache-2.0
language:
- en
pipeline_tag: image-text-to-text
library_name: transformers
---

# **oMEGA-4B-SpatialThink-0804-GGUF**

> **oMEGA-4B-SpatialThink-0804** is a vision-language model built on top of **Qwen/Qwen3-VL-4B-Instruct** and fine-tuned for **spatial reasoning with concise notes for unfiltered vision tasks**. The model is trained to produce concise yet informative reasoning for spatial understanding while maintaining strong image captioning capabilities. Training is based on **remyxai's SpaceThinker** and **OpenCaption-FineGrained**, enabling efficient spatial reasoning and detailed image understanding across diverse visual domains.

> [!NOTE]
> This model is an experimental release and may generate unexpected behaviors or reasoning artifacts in certain scenarios.

## Model Files

   File Name | Quant Type | File Size | File Link |
 |-----------|------------|-----------|-----------|
 | oMEGA-4B-SpatialThink-0804.BF16.gguf | BF16 | 8.05 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.BF16.gguf) |
 | oMEGA-4B-SpatialThink-0804.F16.gguf | F16 | 8.05 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.F16.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q3_K_L.gguf | Q3_K_L | 2.24 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q3_K_L.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q3_K_M.gguf | Q3_K_M | 2.08 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q3_K_M.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q3_K_S.gguf | Q3_K_S | 1.89 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q3_K_S.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q4_K_M.gguf | Q4_K_M | 2.5 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q4_K_M.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q4_K_S.gguf | Q4_K_S | 2.38 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q4_K_S.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q5_K_M.gguf | Q5_K_M | 2.89 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q5_K_M.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q5_K_S.gguf | Q5_K_S | 2.82 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q5_K_S.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q6_K.gguf | Q6_K | 3.31 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q6_K.gguf) |
 | oMEGA-4B-SpatialThink-0804.Q8_0.gguf | Q8_0 | 4.28 GB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.Q8_0.gguf) |
 | oMEGA-4B-SpatialThink-0804.mmproj-bf16.gguf | mmproj-bf16 | 839 MB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.mmproj-bf16.gguf) |
 | oMEGA-4B-SpatialThink-0804.mmproj-f16.gguf | mmproj-f16 | 839 MB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.mmproj-f16.gguf) |
 | oMEGA-4B-SpatialThink-0804.mmproj-q8_0.gguf | mmproj-q8_0 | 454 MB | [Download](https://huggingface.co/prithivMLmods/oMEGA-4B-SpatialThink-0804-GGUF/blob/main/oMEGA-4B-SpatialThink-0804.mmproj-q8_0.gguf) |

## llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp