Community
Reddit user asks whether Gemma 4 E4B is still the best small local tool-calling model
A Mac user running a local assistant with MLX seeks small models under 8B that beat Gemma 4 E4B on tool calling, tool chaining, long synthesis and multimodal input. Ling 3.0 Tiny, LFM models and Qwen 3.5 4B/9B have not worked for them so far.