Qwen API relay services 25 in total

Qwen is the most widely integrated Chinese family among relays, for practical reasons: the vendor's open platform is friendly to individuals and small teams, integration is simple, and the line spans every tier from lightweight to flagship — a cheap small model for bulk jobs and a flagship for complex reasoning, switched through the same API, which makes the cost structure very flexible.

On domestic relays Qwen is usually the cheapest option available. If your work is mainly Chinese-language processing, information extraction and format conversion, a lightweight tier is normally enough, at an order of magnitude below an overseas model of the same tier.

Watch out for tier alignment. Qwen's naming is dense — different parameter sizes and different generation releases are mixed together — and comparing a lightweight tier against another vendor's flagship tells you nothing. Our official price list gives reference prices per tier, so locate your tier first and then compare across.

If your work involves data compliance, the data path matters too: a request through a relay means the content is visible to a third party. For sensitive workloads, go directly to the vendor's open platform or to a gateway you host yourself.

• Most stable：Highest current uptime. Where samples are thin we show the public-dataset baseline instead, and the detail page states the source.
• Fastest：Home-page response time from the most recent probe. A single measurement is noisy, so treat it as a range rather than an absolute.
• Broadest model coverage：This uses the model count our endpoint probe actually detected, not the support list the service advertises — the two are measured differently, and the advertised figure is usually the more optimistic one.

What is Qwen good at?

Chinese information extraction, text classification, format conversion and summarisation — structured work where it is both steady and cheap. It handles code completion too, though it trails flagship models on complex architecture design. The common pattern is tiering: routine work on a lightweight tier, hard problems switched to a flagship.

Why does Qwen cost so much more on some services than others?

Two reasons. First, different tiers are being quoted — lightweight and flagship prices are far apart. Second, the paths differ: a service connected directly to the vendor has costs close to the official ones, while a reseller may be cheaper but with less predictable stability and a murkier data path.

What about data security with Qwen?

A request through a relay means the content is visible to a third party. Where trade secrets or personal data are involved, go through the vendor's official platform or a self-hosted gateway; if you do use a relay, confirm its data-handling and retention terms first.

There are so many Qwen models — how do I choose?

Choose by task complexity, not by how new the model is. Run a batch of real tasks on a lightweight tier and measure accuracy; move only the categories that fall short onto a larger tier. That is far cheaper than starting on a flagship.