Nabbers1999/MS-24B-Bathory-GRPO
TEXT GENERATIONConcurrency Cost:2Model Size:24BQuant:FP8Ctx Length:32kPublished:Mar 21, 2026Architecture:Transformer0.0K Cold

Nabbers1999/MS-24B-Bathory-GRPO is a 24 billion parameter Mistral-based language model, fine-tuned for enhanced writing capabilities and instruction following. It incorporates GRPO (Generative Reinforcement Learning from Prose Optimization) training to refine writing style, reduce AI-generated 'slop', and encourage a prose limit of 768 tokens per response. This model excels in long-context multi-round chat and roleplay, making it particularly suitable for creative writing and interactive narrative generation.

Loading preview...