24B-Suite/Mergedonia-PROMETHEUS-24B-v1

Hugging Face
TEXT GENERATIONConcurrency Cost:2Model Size:24BQuant:FP8Ctx Length:32kPublished:Apr 17, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

Mergedonia-PROMETHEUS-24B-v1 is a 24 billion parameter Mistral-based causal language model developed by 24B-Suite, utilizing a highly experimental 6-stage merge method called PROMETHEUS. This model integrates multiple base models, including TheDrummer's Precog, Magidonia, and Cydonia, through a complex merging pipeline designed for advanced parameter optimization. It is engineered to combine diverse model strengths, making it suitable for general-purpose text generation and complex reasoning tasks.

Loading preview...

Mergedonia PROMETHEUS 24B v1: An Experimental Merge

Mergedonia PROMETHEUS 24B v1 is a 24 billion parameter language model built on the MistralForCausalLM architecture. Developed by 24B-Suite, this model introduces a novel and highly experimental 6-stage merge method, dubbed "PROMETHEUS (Version P)," which combines several official and custom techniques.

Key Capabilities & Features

  • Advanced Merging Pipeline: Utilizes a complex 6-stage pipeline incorporating methods like CABS + RAM Core Settings, Quantum Frankenstein Glue + Shannon Entropy, AETHERIC SYMPLECTIC GROUNDING, and FLUX Frankenstein Surgery.
  • Multi-Model Integration: Merges contributions from several base models, including TheDrummer/Precog-24B-v1, TheDrummer/Magidonia-24B-v4.3, TheDrummer/Cydonia-24B-v4.3, and !BeaverAI_Fallen-Mistral-Small-3.1-24B-v1e_textonly.
  • Configurable Parameters: Features a wide array of configurable parameters for fine-tuning the merge process, such as ram_r, della_eps, iota, zeta, asg_resonance, and phi_optim.
  • Context Length: Supports a context length of 32768 tokens, enabling processing of longer inputs.

Good For

  • General-Purpose Text Generation: Suitable for a broad range of language understanding and generation tasks.
  • Complex Reasoning: The sophisticated merging approach aims to enhance the model's ability to handle intricate prompts and generate coherent responses.
  • Experimentation: Ideal for researchers and developers interested in exploring advanced model merging techniques and their impact on performance.

Finetunes
2 models