Qwen3 30B A3B

Qwen3 MoE: 30B total, 3B active. Tokens per second of a 3B model at the quality of a 30B.

Parameters

30.5B

Family

Alibaba

Context

256K tokens

FP16 weights

61GB

// where you can run it

// hugging face stats (cached daily)

1.1M downloads · 809 likes · license: apache-2.0 · updated 8 months ago

What you need to run this.

$ ./vrambudget --model qwen-3-30b-a3b --by quant

// budgets shown at ctx 8K, concurrency 1, 15% safety headroom. Tune in the calculator →

$ grep --params similar catalog.json

$ gh discussion list

// sign in with github to leave a comment. threads live in the repo's discussions tab.