gemma4 e2b is really good, what other small models work on crappy computers?

I run it on i5 6500 and I get 9t/s its really fast and the output is a lot better than ChatGPT 3.5 and maybe its as good as ChatGPT 4 but I didn't use that 4.0 much. What are other good small models? I used Qwen 3.5 4b before this and that one blew me away too.