Gemma-4-E2B-IT seems to be as good or better than Qwen3.5-4B while having massively shorter reasoning times on average

r/LocalLLaMA
Open Source AI

Submitted by /u/ZootAllures9111 [link] [comments]