How long can I expect to wait until the local ~30B A3B frontier catches up to GLM 5.3 Flash quality?

A user is asking about the timeline for local AI models, specifically around 30B A3B, to reach the quality of GLM 5.3 Flash. They note a significant improvement from Qwen3 Coder 30B A3B to Qwen 3.6 35B A3B in about six months. The user wonders if a year is a reasonable estimate for local AI to eliminate the need for cloud services, given their current hardware of 16 GB RAM and 8 GB VRAM.