Open-weight & challenger labs
Alibaba (Qwen)
One of the most prolific open-weight labs — dozens of model sizes, most free to self-host.
Checked September 19, 2026
No lab publishes a real roadmap
So this page keeps three things apart on purpose. What has already shipped is dated. What is coming is listed only where the lab itself has said it in public, with the source named. Everything else sits under “What to watch” and is my own read of their shipping pattern — informed guessing, not news. If a date matters to you, check the lab's own page linked above.
What they are
Qwen is the most prolific open-weight family on this page — dozens of sizes, from models that run on a laptop up to frontier-scale coding models, most under Apache 2.0.
The flagship is the exception: it is API-only and paid. Everything a beginner would actually reach for is free.
What they currently ship
Every model in the directory entry, with its free-tier status.
- Qwen3.8-Max— Flagship, best for autonomous coding on complex projects; API-only.Paid only
- Qwen3-Coder-480B and Qwen3-235B-A22B— Top open-weight coding and general models, Apache 2.0 licensed.Free, open-weight
- Qwen3 4B / 1.7B / 0.6B— Smaller models that run on a laptop.Free, open-weight
- Note: There is also a free consumer chat app at chat.qwen.ai.
What you get for free
The honest version, in one paragraph.
Very strong. A free chat app, and an open-weight catalog deep enough that most people never need the paid flagship. The laptop-sized models are the easiest genuine entry point to running AI yourself on this whole page.
Track record
Things that actually happened, oldest first, each with a date.
- 2023The first Qwen models publish with open weights and broad language coverage.
- 2024–2025The Qwen2 and Qwen3 families expand to cover coding, vision and very small on-device sizes.
- 2026Qwen3.8-Max leads as a paid API-only flagship, with Qwen3-Coder-480B and Qwen3-235B-A22B as the open-weight heavyweights.
Confirmed and announced
Only what the lab has stated in public. Nothing here rests on a rumor.
Qwen3-Coder-480B, Qwen3-235B-A22B and the small Qwen3 4B / 1.7B / 0.6B models are Apache 2.0 licensed and free to self-host.
Announced by the lab
There is a free consumer chat app at chat.qwen.ai; Qwen3.8-Max is paid and API-only.
Stated on the lab's own pricing or docs page
What to watch
My own read of where this lab is heading, based on how it has shipped so far. Not announcements — treat these as questions to check, not facts.
- — Whether coding-specific open-weight releases keep pace with the paid flagship, which is where Qwen has been strongest.
- — The small-model line, which is the most realistic route to running something capable on a laptop.
- — License terms per release — Qwen is usually Apache 2.0, but check the specific model rather than assuming.
Their ethics and safety policies
What Alibaba (Qwen) has published about safety and ethics, the documents themselves, what those documents leave out — and my own read, kept separate.
Looking for which tool to actually use? See AI Models Compared