Qwen3.5 Small

Compact multimodal language models from 0.8B to 9B parameters

Visit Qwen3.5 Small →

Qwen3.5 Small is a series of open-weight language models (0.8B, 2B, 4B, 9B) with native multimodal capabilities and reinforcement learning improvements. The smaller variants (0.8B, 2B) target edge deployment, while the 4B and 9B models serve as lightweight agent bases competitive with larger models. Designed for developers needing efficient, deployable AI across constrained hardware environments.

At a glance

Company
Alibaba Cloud (Qwen)
Pricing
free
API available
Yes
Self-hostable
Yes
Launched
2026-03
Last verified
2026-05-11

Capabilities

multimodaledge-deploymentfine-tuninglong-contextopen-weightsrl-training

Categories

Alternatives

For AI agents: machine-readable markdown version of this page at /tools/qwen3-5-small.md, or send Accept: text/markdown.