✉ The Friday AI Brief: the week's 5 best AI stories, tools & comparisons — in your inbox every Friday morning.

Alibaba Group logo

Alibaba Launches Qwen4-Max for Real: Full Release Brings Pricing, a 1M Context and an API Anyone Can Use

Alibaba has fully launched Qwen4-Max, turning the flagship model it previewed on November 11 into a generally available release with published API pricing, a one-million-token context window and native vision. Why it matters: this is the launch that counts — the preview disappears on November 25, but Qwen4-Max itself is now a product anyone can call, at prices that undercut most Western flagships.

November 16, 2026

What changed from the preview

When Alibaba opened Qwen4-Max in preview last week, access was time-boxed and the commercial details were missing. The full launch fills them in. Qwen4-Max is a sparse mixture-of-experts model built on a 2.4-trillion-parameter backbone, with native visual understanding and support for a context window of one million tokens. It is available through the API on QwenCloud, with OpenAI- and Anthropic-compatible protocols, and has been listed on OpenRouter.

Pricing is set at $1.50 per million input tokens and $4.50 per million output tokens, with cached input at $0.075 per million. The flagship itself stays closed-weight, but Alibaba says smaller open-weight Qwen4 models are coming — including a Qwen4-27B the company says will run on machines with 16GB of RAM.

The performance claims

Alibaba says Qwen4-Max scored 75.7 on the DeepSWE v1.1 coding benchmark, ahead of Anthropic’s Claude Opus 5.5, and ranked sixth in CodeArena’s front-end coding blind test. Independent tracker Artificial Analysis puts it at 43 on its Intelligence Index, nine points above its predecessor Qwen3.8-Max. The benchmark scores are the company’s own reported results; the Artificial Analysis figure is third-party.

The launch also came with an unusually prominent safety disclosure. Alibaba said Qwen4-Max’s training included security checks alongside coding tests, aimed at reducing vulnerabilities such as code injection and broken access controls, and that safeguards against misuse in areas like cybercrime and hazardous materials were strengthened. Internal testing found no evidence of significant high-risk capabilities, the company added — a claim Reuters noted fits a broader pattern of Chinese labs expanding safety disclosures in new releases.

Why it matters

China’s labs are no longer just winning on open weights; they are shipping closed flagships with transparent pricing and going straight at US rivals on coding and professional work. For subscribers, the practical read is that a top-tier coding model now costs $1.50/$4.50 per million tokens from a major provider — real pressure on every Western price list. Everything AI, tested: the preview was a teaser; this launch, with pricing and availability anyone can verify, is the story.

Frequently asked questions

How much does Qwen4-Max cost?

Alibaba prices it at $1.50 per million input tokens and $4.50 per million output tokens on QwenCloud, with cached input at $0.075 per million tokens.

Is Qwen4-Max open source?

No. The flagship remains closed-weight. Alibaba says smaller open-weight Qwen4 models are coming soon, including a 27B model designed to run on 16GB of RAM.

How is this different from the November 11 preview?

The preview offered time-limited access that ends November 25. The full launch adds published pricing, general API availability on QwenCloud and an OpenRouter listing.

Sources: Reuters, Bloomberg via The Hindu Business Line, Longbridge (Wallstreetcn), KuCoin

Leave a Comment

Your email address will not be published. Required fields are marked *

Get the 5 best AI tools every week

Top AI news, tools, and prompts — one short email. Free, unsubscribe anytime.

Run a newsletter of your own? Monetize and grow it with SparkLoop →

Scroll to Top