ApocalypseParty/G4-31B-it-base10

VISIONConcurrency Cost:2Model Size:31BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jun 24, 2026Architecture:Transformer Cold

ApocalypseParty/G4-31B-it-base10 is a 31 billion parameter language model created by ApocalypseParty, merged from Google's Gemma-4-31B-it and Gemma-4-31B models using the SLERP method. This model leverages the instruction-tuned capabilities of Gemma-4-31B-it combined with the base Gemma-4-31B model, offering a 32768 token context length. It is designed for general language understanding and generation tasks, building upon the strengths of its constituent Gemma models.

Loading preview...

Model Overview

ApocalypseParty/G4-31B-it-base10 is a 31 billion parameter language model developed by ApocalypseParty. It is a product of a strategic merge of two foundational models from Google: google/gemma-4-31B-it and google/gemma-4-31B. This merge was executed using the SLERP (Spherical Linear Interpolation) method, a technique often employed to combine the strengths of different pre-trained models while maintaining performance.

Key Characteristics

  • Parameter Count: 31 billion parameters, providing substantial capacity for complex language tasks.
  • Context Length: Supports a context window of 32768 tokens, enabling processing of longer inputs and generating more coherent, extended outputs.
  • Merge Method: Utilizes the SLERP merge method, which aims to blend model weights effectively, potentially enhancing overall performance or specializing capabilities.
  • Base Models: Built upon Google's Gemma-4-31B-it (instruction-tuned) and Gemma-4-31B (base model), inheriting their respective strengths in instruction following and foundational language understanding.

Potential Use Cases

This model is well-suited for applications requiring:

  • General-purpose text generation: Creating diverse forms of content, from creative writing to informative summaries.
  • Instruction following: Benefiting from the instruction-tuned component for tasks requiring specific output formats or responses to direct commands.
  • Long-context understanding: Handling and generating text for scenarios where extended conversational history or document analysis is crucial.

Finetunes
1 models