ZeroXClem/Gemma3-4B-Arceus-Servant

VISIONConcurrency Cost:1Model Size:4.3BQuant:BF16Ctx Length:32kPublished:Nov 9, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

ZeroXClem/Gemma3-4B-Arceus-Servant is a 4.3 billion parameter Gemma 3 model created by ZeroXClem, built upon mlabonne/gemma-3-4b-it-abliterated. This composite model, formed via MODEL STOCK merging of six distinct Gemma3-borne personalities, excels as a balanced omnidirectional companion. It is optimized for tactical reasoning, competitive programming, Claude-style creative writing, and high instruction compliance, supporting a context length of up to 262,144 tokens.

Loading preview...

ZeroXClem-Gemma3-4B-Arceus-Servant: An Elite Composite Model

ZeroXClem-Gemma3-4B-Arceus-Servant is a 4.3 billion parameter Gemma 3 model, developed by ZeroXClem, that leverages the mlabonne/gemma-3-4b-it-abliterated core. This model is a "celestial convergence" achieved through MODEL STOCK merging, integrating six different Gemma3-based models to create a versatile and powerful AI.

Key Capabilities

  • Uncensored Expression: Inherits capabilities from gemma-3-4b-it-abliterated for unrestricted generation.
  • Claude-Style Eloquence: Provides natural, thoughtful, and creative text generation.
  • Tactical Intelligence: Excels in constraint-aware reasoning, offering smart breakdowns and clarification prompts.
  • Code Competence: Optimized for competitive programming and structured logic tasks.
  • Creative Persona: Combines warmth, storytelling, and multilingual depth.
  • High Instruction Compliance: Designed to follow commands cleanly and precisely.

Technical Specifications

Built on a 34-layer Gemma 3 4B architecture, the model supports bfloat16 precision and boasts an impressive context length of up to 262,144 tokens via rope scaling. It is compatible with GGUF and 4-bit inference for efficient deployment.

Good For

This model is ideal for use cases requiring a balanced omnidirectional companion, including complex reasoning tasks, code generation and explanation, creative writing, and scenarios demanding high instruction adherence and meta-cognitive decision-making.