About QwenAPI.net

A topic site tracking Qwen model API access — starting with Qwen3.8 Max, whose launch week has produced more confusion than facts.

Who runs this

QwenAPI.net is operated by the team behind EvoLink, a unified pay-as-you-go API gateway for 50+ AI models. That's a commercial interest, and we'd rather state it than have you discover it in a footer scan: when we list ways to call Qwen models, EvoLink appears first, with the reasons given. We are not affiliated with Alibaba Cloud, and we are not an independent testing lab.

Editorial rules

• Every number is sourced and dated. Model facts are verified against official documentation on the stated date, not copied from news coverage — we've already caught widely-repeated pricing errors that way.
• Vendor claims are labeled as claims. "2.4T parameters" and "second only to Fable 5" stay tagged as unverified until a model card or benchmark table exists.
• Unknowns stay unknown. Active parameters, PAYG pricing, license, GA dates: not announced means we write "not announced", not a guess.
• No unexecuted code. We publish code samples only after running them ourselves, verbatim. Until a public API exists, that means no code samples.
• Evaluations disclose their conditions. When our coding-agent evaluation results publish, they come with tasks, dates, run counts, and limitations attached.

Contact

Corrections are welcome and get priority handling — if a fact on this site is wrong or stale, tell us via the contact channels on EvoLink.