yamatazen/Himeyuri-Magnum-12B-HereticLoRA

TEXT GENERATIONPricing:Input $0.87 / Cached $0.2 / Output $0.99Concurrent Unit Cost:1Model Size:12BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Sep 8, 2026Architecture:Transformer0.0K Featherless Exclusive Cold

yamatazen/Himeyuri-Magnum-12B-HereticLoRA is a 12 billion parameter ChatML model created by yamatazen, merged using the Model Stock method with Elizezen/Himeyuri-v0.1-12B as its base. This model integrates components from anthracite-org/magnum-v2-12b and anthracite-org/magnum-v2.5-12b-kto, offering a combined capability for conversational AI tasks. It is designed for applications requiring a robust 12B parameter language model with a 32768 token context length.

Loading preview...

Model Overview

yamatazen/Himeyuri-Magnum-12B-HereticLoRA is a 12 billion parameter language model, specifically configured as a ChatML model. It was developed by yamatazen through a sophisticated merging process using mergekit and the Model Stock merge method.

Key Characteristics

  • Base Model: The foundation of this model is Elizezen/Himeyuri-v0.1-12B.
  • Merged Components: It incorporates capabilities from two distinct models:
  • Merge Method: Utilizes the Model Stock method, ensuring a structured integration of the merged models.
  • Data Type: The model was processed using bfloat16 for efficiency.
  • Context Length: Supports a substantial context window of 32768 tokens.

Intended Use Cases

This model is well-suited for conversational AI applications due to its ChatML format. Its merged architecture suggests a blend of capabilities from its constituent models, making it potentially versatile for various text generation and understanding tasks where a 12B parameter model with a large context window is beneficial.

Finetunes
2 models