Liquid Says Its 1 GB Model Beats Qwen’s 9B. We Tested Both

Liquid AI released LFM3 on October 27, 2026: six 2B-class open-weight models. Liquid says: "Built for the edge: the text model fits in about 1 GB at 4-bit precision." On the average row of Liquid's own benchmark table, LFM3-1.7B scores 49.37 against 47.24 for Qwen3.5-9B, a model more than five times its size. Liquid also writes: "Larger models still win on agentic coding and knowledge, where size matters most." Its Hugging Face model card says: "It is not recommended for agentic coding or complex multi-step reasoning due to its small size."