Atlas-labs/atlas-axiom-1.5b

Atlas-Axiom-1.5b (Everything Mixture)

Axiom-1.5b is a high-performance generalist reasoning model developed by Atlas Labs.

It has been fine-tuned on the Everything Mixture, a diverse dataset covering:

• —General Reasoning & Logic
• —Science & History
• —Programming (Python, JS, etc.)
• —Creative Writing
• —Casual Conversational Intelligence

Model Details

• —Developer: Atlas Labs
• —Base Model: Qwen2.5-1.5B-Instruct
• —Architecture: Causal Language Model
• —Training Method: SFT (Supervised Fine-Tuning) with LoRA

Usage

This model is designed for use in the Axiom Chat Space.