---
license: apache-2.0
base_model:
- ertghiu256/qwen3-4b-code-reasoning
- ValiantLabs/Qwen3-4B-Esper3
- Tesslate/UIGEN-T3-4B-Preview
- Qwen/Qwen3-4B
tags:
- mergekit
- ties
- thinking
- reasoning
- merge
- code
---
# Model Info
Based on the Qwen 3 4b parameter model. This model is merged from models that are trained on reasoning coding tasks.

# Use cases
- Coding (HTML, C++, Python)
- Agentic coding
- General assistant

# Drawbacks
- Overly Overthinking

# Recommended Parameters
- context length: >= 8000 
- temperature: 0.6
- top p: 40
- top k: 0.95