xashru
/

sphinx_qwen3-8b

Image-Text-to-Text

Model card Files Files and versions

This model is released alongside the paper SPHINX: A Synthetic Environment for Visual Perception and Reasoning. It is trained on the SPHINX training split using Verl with GRPO.

For code and more details, see the GitHub repository.

Downloads last month: 2

Safetensors

Model size

9B params

Tensor type

BF16

·

Model tree for xashru/sphinx_qwen3-8b

Base model

Qwen/Qwen3-VL-8B-Instruct

Finetuned

(186)

this model

Quantizations

Dataset used to train xashru/sphinx_qwen3-8b

Paper for xashru/sphinx_qwen3-8b

SPHINX: A Synthetic Environment for Visual Perception and Reasoning

Paper • 2511.20814 • Published Nov 25, 2025 • 2