htdevs/ phinegpt:latest

16 2 days ago

PhineGPT is a custom AI assistant model built on top of Gemma4 31B using Ollama. It is designed as a general-purpose assistant optimized for accuracy, structured reasoning, and developer-friendly output.

vision tools thinking
ollama run htdevs/phinegpt

Details

2 days ago

c18f092859ad · 20GB ·

gemma4
·
31.3B
·
Q4_K_M
Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR US
You are PhineGPT, a highly capable general-purpose AI assistant. Core principles: - Be accurate and
PhineGPT License Copyright (c) 2026 PhineGPT Permission is hereby granted, free of charge, to any pe
{ "num_ctx": 4096, "repeat_last_n": 512, "repeat_penalty": 1.12, "temperature": 0.3,

Readme

PhineGPT

PhineGPT is a custom AI assistant model built on top of gemma4:31b using Ollama. It is designed as a general-purpose assistant optimized for accuracy, structured reasoning, and developer-friendly output.


Model Overview

PhineGPT is intended for local execution via Ollama and provides a reliable assistant for:

  • Software development and debugging
  • Mathematics and logical reasoning
  • Hardware and embedded systems
  • General knowledge and writing assistance

The model prioritizes correctness, clarity, and structured responses over verbosity or speculation.


Model Behavior

PhineGPT follows these principles:

  • Prioritizes correctness over speed or verbosity
  • Avoids hallucinating facts, APIs, or external capabilities
  • Requests clarification when input is ambiguous
  • Produces structured and readable responses
  • Adapts tone to match the user while remaining professional

Model Parameters

  • temperature: 0.3
  • top_p: 0.9
  • repeat_penalty: 1.12
  • num_ctx: 8192
  • repeat_last_n: 256

These settings balance reasoning stability, coherence, and controlled creativity.


License

This model is distributed under the PhineGPT License.

  • Attribution is required for any usage or redistribution
  • Modified versions must include attribution and a modification notice
  • Provided as-is without warranty

Author

PhineGPT


Notes

  • Optimized for local inference using Ollama
  • Best performance achieved on GPU-accelerated systems
  • Suitable for development, education, and general assistance tasks