Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Vocabulary size ~248k. A bit bigger than other recent Chinese models (Kimi K3 ~164k, DeepSeek-V4 ~129k, and GLM-5.2 ~155k).

Make of this what you will.



> Make of this what you will.

I'm interested in your take on it. IIRC Gemma family models too have a ~250k vocabulary size


Does this mean its tokenizer is somehow tuned?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: