# Qwen 3.5 Small — Apache Multimodal Edge From 0.8B to 9B

Times of AI Desk · 2026-03-01 · Models

[https://timesof.ai/2026/03/alibaba-releases-qwen-3-5-small-series-for-on-device-ai](https://timesof.ai/2026/03/alibaba-releases-qwen-3-5-small-series-for-on-device-ai)

> Alibaba Qwen completed Qwen 3.5 rollout with Small series (March 1): dense 0.8B/2B/4B/9B open weights, native text/image/video in one stack, Apache 2.0 — on-device/edge after mid-February 397B MoE flagship. 9B laptop-class; smaller for phones. Competitive-vs-larger benches are vendor/community — not Artificial Analysis locks.

U.S. labs chase cluster Elo; Alibaba’s proprietary close of the Qwen 3.5 family is **ship the edge SKUs** — native multimodal dense models small enough for phones and laptops under **Apache 2.0**.

**Alibaba Qwen** released the **Qwen 3.5 Small** series (March 1): four dense open-weight models — **Qwen3.5-0.8B**, **2B**, **4B**, **9B** — optimized for local inference on smartphones, laptops, and resource-constrained hardware. Native multimodal (text, image, video) in the same weights without separate vision adapters. Builds on larger Qwen 3.5 foundation that began with a **397B MoE** flagship in mid-February. Developers: 9B on consumer laptops; smaller variants on phones/older hardware. Reports: 9B competitive vs much larger models on some targeted benches (community/vendor). License: **Apache 2.0**. Context: qwen.ai/blog family announcements.

## Claims vs checks

Model sizes, Apache 2.0, and on-device positioning are **Qwen/Alibaba primary** plus early-March coverage. Bench “competitive with larger” claims are **vendor/community** — verify on Artificial Analysis or task-specific boards before ranking freezes. Original frontmatter erroneously listed `google` as platform; omitted here.

## Limits

- On-device quality depends on quantization and NPU/GPU path.
- Small multimodal still hallucinates; offline ≠ private if apps phone home.
- MoE flagship and Small series serve different buyers — do not conflate.

## Sources

- [Qwen blog / Qwen3.5 announcement context](https://qwen.ai/blog).
- Alibaba Qwen announcements on model family rollout.
- Benchmark and deployment analysis from developer communities (March 2026).