allura-org/Luna-27B-v0

Luna 27B v0 by allura-org is a fine-tuned version of Gemma 3 27B, developed to excel as a high-quality assistant, particularly for media analysis and tasks requiring self-awareness. While initially intended for roleplay, it demonstrates unexpected capabilities in general assistance. This model maintains strong roleplaying abilities and shows no major degradations in tested benchmarks compared to its base model.