Model Watch · Alibaba
Qwen3
4 models carry the Qwen3 name, 4 of them callable today. Here's what separates them, and what each one costs.
From Alibaba's own docs · updated 2026-09-05 10:32 UTC
Qwen3.8-27B shipped on 2026-08-05. Running one of these for a month costs anything from $0.27 on Qwen3.8-Flash to $3.52 on Qwen3.8-Max, for identical work.
Every Qwen3 model
| Model | Status | Released | Context | In / 1M | Out / 1M | Your month |
|---|---|---|---|---|---|---|
| Qwen3.8-27B | released | 2026-08-05 | — | — | — | — |
| Qwen3.8-Max | released | — | 1M | $2.00 | $6.00 | $3.52 |
| Qwen3.8-Flash | released | — | 1M | $0.15 | $0.47 | $0.27 |
| Qwen3.7-Plus | released | — | 1M | $0.40 | $1.60 | $0.86 |
The highlighted row is the newest one Alibaba has actually released. The monthly figure is 700 words in and 500 out, twenty times a day, at list prices. Change the workload.
Which one should you be using?
Qwen3.8-27B is the newest released model in the family, out on 2026-08-05. Newer is not automatically better for your job, but it is usually the one the lab is optimising, and on this table it is rarely the most expensive.
Why there are so many numbers
A point release is a new model, not a patch. Alibaba ships a new one under the same family name when the training or the tuning changes, and the old one usually keeps working at its old price until it's deprecated.
That's why two models in this table can hold the same amount of text and cost different amounts. You're picking a specific model, not a family, and the API name in each model's page is what actually goes in your code.
What each one is for
- Qwen3.8-27BThe dense open-weight member of the 3.8 family. No API price because you host it yourself. The date is when the repo went up.
- Qwen3.8-MaxAlibaba's flagship, at a third of what Fable 5.1 costs. Prices are region-scoped: this is the Singapore column, and Beijing is cheaper again.
- Qwen3.8-FlashFifteen cents a million with a million tokens of context. The cheapest million-token window in this whole catalogue.
- Qwen3.7-PlusThe mid tier. Prompts over 256K tokens triple the input rate, so the sticker price only holds for short calls.
Keep reading
Alibaba prices in full · find a model by budget · compare any two