OneThinker-8B-AIO-GGUF

The OneThinker-8B model, including the OneThinker-SFT-Qwen3-8B variant, is an all-in-one reasoning model for images and videos based on the Qwen-3-VL-Instruct-8B architecture, optimized for comprehensive multimodal reasoning tasks such as image question answering, video understanding, grounding, tracking, and segmentation. It achieves top-tier performance across a wide range of benchmarks including 70.6% accuracy on MMMU for image QA and strong video QA results, surpassing other leading open-source models like Qwen3-VL-Instruct-8B. The model uses a unified text format to integrate diverse reasoning tasks, enabling advanced STEM, general knowledge, and multimodal reasoning capabilities. It supports large context lengths and has been trained with extensive GPU resources, making it a powerful tool for visual and video reasoning tasks with state-of-the-art results in multimedia understanding and question answering.

OneThinker-8B [GGUF]

File Name	Quant Type	File Size	File Link
OneThinker-8B.BF16.gguf	BF16	16.4 GB	Download
OneThinker-8B.F32.gguf	F32	32.8 GB	Download
OneThinker-8B.Q8_0.gguf	Q8_0	8.71 GB	Download
OneThinker-8B.mmproj-bf16.gguf	mmproj-bf16	1.16 GB	Download
OneThinker-8B.mmproj-f32.gguf	mmproj-f32	2.31 GB	Download
OneThinker-8B.mmproj-q8_0.gguf	mmproj-q8_0	752 MB	Download

OneThinker-SFT-Qwen3-8B [GGUF]

File Name	Quant Type	File Size	File Link
OneThinker-SFT-Qwen3-8B.BF16.gguf	BF16	16.4 GB	Download
OneThinker-SFT-Qwen3-8B.F32.gguf	F32	32.8 GB	Download
OneThinker-SFT-Qwen3-8B.Q8_0.gguf	Q8_0	8.71 GB	Download
OneThinker-SFT-Qwen3-8B.mmproj-bf16.gguf	mmproj-bf16	1.16 GB	Download
OneThinker-SFT-Qwen3-8B.mmproj-f32.gguf	mmproj-f32	2.31 GB	Download
OneThinker-SFT-Qwen3-8B.mmproj-q8_0.gguf	mmproj-q8_0	752 MB	Download

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):