Qwen2.5-1.5B-Instruct

Qwen2.5-1.5B-Instruct is a compact 1.54-billion-parameter instruction-tuned language model from Alibaba's Qwen team, designed for general-purpose text generation, multilingual conversation, coding, and math tasks. It supports structured outputs and a 128K-token context window, making it well-suited for lightweight assistant and developer workflows. Source: supabase.