AI & Tech··3 sources verified
Alibaba Launches Qwen4-Max, Claims 62% Lower Cost Per Task and Adds Cyber Safeguards
Summarised by Relevant News AI · Read time: 3 min
Alibaba released Qwen4-Max, a 2.4-trillion-parameter mixture-of-experts model with a 1-million-token context window, which it says cuts the average cost per professional task to $1.67 from $4.43 for its predecessor. The savings come from fewer tokens used and lower list prices of $1.50 per million input and $4.50 per million output tokens. The company also added safeguards against offensive cybersecurity capabilities and other high-risk uses, though the cost figures come from its own proprietary Work360 benchmark and analysts caution they are not guaranteed enterprise savings.
Why it matters: Teams evaluating frontier models should weigh vendor cost-per-task claims against their own cost of accepted results, and note the new cyber-use restrictions when planning deployments.