CG@cgtwts
Qwen 3.5 Small models
- fully open source
- beats models 4x it’s size
- 9B model performs on par with GPT OSS 120B while being 13x smaller
- outperforms Gemini 3 flash and Claude sonnet 4.5 on select benchmarks
- runs on any laptop
- even works on a phone
- completely free.
Opens with an observation
