TheDrummer/Orion-26B-A4B-v1.1

Hugging Face
VISIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:2Model Size:26BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 13, 2026Architecture:Transformer0.1K Featherless Exclusive Warm

TheDrummer/Orion-26B-A4B-v1.1 is a 26 billion parameter language model developed by TheDrummer, featuring a 32768-token context length. This model is specifically fine-tuned for creativity, usability, and entertainment, prioritizing dynamic storytelling and imaginative text generation over strict intelligence or problem-solving. It aims to enhance user experience in creative applications, offering a distinct approach compared to models optimized for non-creative uses.

Loading preview...

Orion 26B A4B v1.1: A Model for Creativity and Entertainment

Developed by TheDrummer, Orion 26B A4B v1.1 is a 26 billion parameter language model with a 32768-token context length, specifically engineered to excel in creative and entertainment-focused applications. Unlike many foundational models optimized for intelligence and problem-solving, Orion prioritizes qualities such as writing style, dynamism, and imagination, aiming to provide a compelling and intriguing storytelling experience.

Key Capabilities

  • Creative Writing: Generates pleasant and effective prose, mimicking a writer's style.
  • Dynamic Storytelling: Excels at crafting compelling and intriguing narratives.
  • Imaginative Generation: Navigates complex possibilities, maintaining coherence while exploring diverse themes.
  • Flexible (Dis)alignment: Designed to avoid corporate/religious/political biases and rigid moralizing, allowing for broader thematic exploration.
  • Instruction Adherence: Capable of following instructions and understanding nuanced prompts.

Good For

  • Creative Use Cases: Ideal for applications requiring imaginative text generation, storytelling, and roleplay.
  • Entertainment: Suited for scenarios where dynamic and engaging content is paramount.
  • Exploration of Diverse Themes: Its design allows for handling difficult or dubious themes without forced positivity or moralizing.

For optimal performance, it is recommended to use a low temperature (e.g., 0.9) and the Gemma 4 31B template. The model's philosophy centers on broadening language perception beyond typical safety and compliance optimizations, offering a unique tool for creative expression.

Finetunes
2 models