cyberlangke/Qwen3-0.6B-Meow-test
TEXT GENERATIONConcurrency Cost:1Model Size:0.8BQuant:BF16Ctx Length:32kPublished:Apr 10, 2026License:mitArchitecture:Transformer0.0K Open Weights Cold

Qwen3-0.6B-Meow-test is an 0.8 billion parameter language model developed by cyberlangke, fine-tuned from Qwen/Qwen3-0.6B-Base and merged into Qwen/Qwen3-0.6B. It supports a context length of 32768 tokens and is characterized by its "cat-like" conversational style, including a tendency to use specific "cat-flavored" responses. While it attempts to support thinking and tool-calling, these features are noted as unstable and prone to hallucination, making its primary use case for generating whimsical, cat-themed dialogue.

Loading preview...

Qwen3-0.6B-Meow-test Overview

Qwen3-0.6B-Meow-test is an 0.8 billion parameter language model developed by cyberlangke. It was created by training a lora adapter on Qwen/Qwen3-0.6B-Base and then merging it into Qwen/Qwen3-0.6B. The model supports a substantial context length of 32768 tokens.

Key Characteristics

  • "Cat-like" Personality: The model is specifically designed to exhibit a "cat-like" conversational style, often referred to as having a "meow flavor" in its responses and chain-of-thought processes.
  • Experimental Features: It includes experimental support for thinking and toolcall, though these functionalities are noted as unstable and may produce hallucinations, particularly with tool invocation results.
  • Multi-turn Dialogue: The model may exhibit repetition in multi-turn conversations.
  • System Prompt: To activate its intended persona, users should employ the system prompt: 你是猫娘奈奈。

Good for

  • Whimsical Chatbots: Ideal for applications requiring a unique, cat-themed conversational agent.
  • Creative Text Generation: Suitable for generating text with a distinct, playful personality.
  • Experimentation: Developers interested in exploring the effects of specific fine-tuning on model persona and experimental features like thinking and toolcall.