Solomon-0.5B is a mobile-focused reasoning model fine-tuned from Qwen/Qwen2.5-0.5B on a curated dataset of Opus CoT reasoning traces.