aimeri/spoomplesmaxx-thrasher-24B

TEXT GENERATIONPricing:Input $0.7 / Cached $0.04 / Output $1.16Concurrent Unit Cost:2Model Size:24BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 27, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The aimeri/spoomplesmaxx-thrasher-24B is a 23.6 billion parameter causal language model, built on Mistral-Small-3.1-24B-Base, specifically fine-tuned for roleplay and creative writing. It features a 32768 token context length and utilizes a custom ChatML template on a Mistral base without vocabulary resizing. This model excels at generating character-driven narratives and is optimized for deployment on single 24GB VRAM cards.

Loading preview...

SpoomplesMaxx-Thrasher-24B: A Roleplay-Focused Model

SpoomplesMaxx-Thrasher-24B is a 23.6 billion parameter model derived from Mistral-Small-3.1-24B-Base, with its vision tower removed and a custom ChatML interface implemented. It is primarily designed for roleplay and creative writing, trained on a corpus that emphasizes character interaction and narrative generation rather than reasoning traces.

Key Capabilities

  • Optimized for Roleplay: Fine-tuned extensively on roleplay data, including the full Toolmaxx family, to produce engaging and in-character responses.
  • Efficient Deployment: At 23.6B parameters, it is designed to be quantized and run on a single 24GB VRAM card, making it accessible for many users.
  • Custom ChatML Template: Features a unique ChatML template that re-purposes existing Mistral tokenizer slots, avoiding vocabulary resizing and ensuring broad compatibility with RP frontends.
  • Tool Competence: Includes corpus-taught conversational tool competence, allowing for tool use within narratives, though not as a strict function-calling API.
  • Extended Context: Supports a context length of 32768 tokens, with the best-trained region up to ~24K tokens.

Good For

  • Character-driven Roleplay: Excels at maintaining character consistency and generating dynamic interactions based on detailed character cards.
  • Creative Writing: Ideal for generating narrative content, dialogue, and descriptive passages.
  • Users with 24GB VRAM: Specifically engineered to run efficiently on hardware with 24GB of VRAM, offering a powerful RP model in a manageable size.

Note: This model is not intended for reasoning, assistant tasks, or safety-critical applications. It is designed to stay in character and does not include refusal mechanisms for out-of-character prompts.