Ame33/Gemma-4-Novelist-Eclipse-31B
VISIONPricing:Input $0.48 / Cached $0.1 / Output $1.44Concurrent Unit Cost:2Model Size:31BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 11, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold
Ame33/Gemma-4-Novelist-Eclipse-31B is a 31 billion parameter language model based on the Gemma-4 architecture, developed by Ame33. This model is a merge of several fine-tuned models, specifically designed to enhance creative writing, focusing on descriptive prose and narrative consistency. With a 32768 token context length, it aims to provide more stable and creative outputs for storytelling applications.
Loading preview...
Overview
Ame33/Gemma-4-Novelist-Eclipse-31B is a 31 billion parameter model built upon the Gemma-4 architecture, developed by Ame33. It is a merged model, evolving from the "Novelist" concept to infuse greater creativity and stability into its outputs. The model incorporates an lm_head with StyleTune for refined text generation.
Key Capabilities
- Enhanced Creative Writing: Specifically designed to improve descriptive prose and narrative flow, building on the original "Novelist" idea.
- Improved Consistency: Addresses previous inconsistencies through a multi-phase merging process, aiming for more coherent storytelling.
- Advanced Merging Strategy: Utilizes a two-phase
mergekitconfiguration, combining multiple base models like G4-Gutenberg, Melinoe, Glimmer, NovelistX, GarnetV2, and Gemopus to achieve its specialized capabilities. - Gemma-4 Base: Leverages the foundational strengths of the Gemma-4 instruction-tuned model from Google DeepMind.
Good For
- Creative Storytelling: Generating detailed and imaginative narratives.
- Descriptive Text Generation: Producing rich, descriptive passages for various writing tasks.
- Roleplay and Interactive Fiction: Creating consistent and engaging character interactions and plot developments.