X-Reasoner-7B-f32-GGUF

X-Reasoner-7B-f32-GGUF

Microsoft X-Reasoner-7B is a 7B-parameter vision-language model from Microsoft Research, post-trained solely on general-domain text using a two-stage approach to achieve generalizable reasoning across modalities and domains without vision-specific fine-tuning. Designed for robust performance on multimodal benchmarks like mathematical reasoning, visual question answering, and cross-domain tasks, it leverages instruction tuning and reinforcement learning to produce structured reasoning traces that transfer zero-shot to unseen visual and textual challenges. The model demonstrates strong generalization by outperforming vision-augmented baselines on held-out domains through text-only post-training, making it particularly valuable for research into modality-agnostic reasoning capabilities under an Apache 2.0 license with Transformers deployment support.

LightOnOCR-2-1B [GGUF]

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):