RynnBrain-Plan-8B-f32-GGUF

RynnBrain-Plan-8B from Alibaba DAMO Academy is an 8B-parameter physics-aware vision-language model serving as an embodied "brain" for robotics, specialized in generating precise manipulation planning instructions by grounding language to physical space/time through spatiotemporal localization (objects, areas, trajectories), egocentric video cognition (QA, counting, OCR), and integration of affordances/objects into downstream VLA outputs for reliable robotic execution. Part of the broader RynnBrain family (2B/8B dense, 30B MoE variants plus Nav/CoP specials), it excels in comprehensive scene understanding from egocentric views, physics-grounded reasoning interleaving text/spatial modalities, and planning for complex tasks in unstructured environments via unified encoder-decoder architecture compatible with ROS 3.0. Released open-source under permissive license with GGUF quantizations, it supports notebooks for OCR/object/area localization and trajectory prediction, positioning Alibaba at the forefront of embodied AI for humanoid robotics.

RynnBrain-Plan-8B [GGUF]

File Name	Quant Type	File Size	File Link
RynnBrain-Plan-8B.BF16.gguf	BF16	16.4 GB	Download
RynnBrain-Plan-8B.F32.gguf	F32	32.8 GB	Download
RynnBrain-Plan-8B.mmproj-bf16.gguf	mmproj-bf16	1.16 GB	Download
RynnBrain-Plan-8B.mmproj-f32.gguf	mmproj-f32	2.31 GB	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):