BlockEdge

Technology and digital life

Alibaba launches Qwen4-Max with 2.4 trillion parameters and a one-million-token context window

Alibaba launched Qwen4-Max on November 16, 2026, according to a report published the following day. The company says the model has 2.4 trillion total parameters, activates 71 billion per token, supports a one-million-token context window and has native visual capabilities.

Qwen4-Max is available through the Qwen AI platform’s API and in Alibaba’s Qwen Office, Qoder and Qwen app. Alibaba says cache-hit pricing starts at 0.5 yuan per million tokens. It reports that the average cost of completing a task is 58% lower than with Qwen3.8-Max, a comparison that should not be read as the price of every task or API request.

Alibaba launches Qwen4-Max with 2.4 trillion parameters and a one-million-token context window

In benchmark results reported by Alibaba, Qwen4-Max ranked sixth for front-end development on CodeArena. At the same Xhigh reasoning setting in Terminal-Bench 4.0, its average token use per attempt fell from 289,000 for Qwen3.8-Max to 151,000, while its score rose from 38.9 to 48.5. These are reported evaluation results, not a guarantee of performance on users’ own projects.

Alibaba also describes uses in manufacturing, finance and scientific research. It says Qwen4 reduced manual programming time for machining a single workpiece from 475 to 24 minutes in an automotive manufacturing scenario. The report does not establish that this result applies across production settings. Smaller models in the Qwen4 series are expected to be released as open source; they had not been described as released in the November 17 report.