Qwen
Alibaba's open-weight model family.
8 models
Qwen2.5 0.5B Instruct
Tiny and instant — runs on laptops without a graphics card
Apache-2.0 · from 398 MB
Qwen2.5 7B Instruct
The previous generation's all-rounder
Apache-2.0 · from 4.7 GB
Qwen2.5 Coder 7B Instruct
Code completion and review, locally
Apache-2.0 · from 4.7 GB
Qwen3 8B
The best all-rounder for an 8GB card
Apache-2.0 · from 5.0 GB
Qwen3 14B
The all-rounder, with more depth
Apache-2.0 · from 9.0 GB
Qwen3.8 27B
The strongest all-rounder you can run on one card
Apache-2.0 · from 16.5 GB
Qwen3 30B A3B
30B of knowledge, 3B of work per word
Apache-2.0 · from 18.6 GB
Qwen3 32B
The largest dense model that fits 24GB
Apache-2.0 · from 19.8 GB