Alibaba's Qwen4-Max takes on Opus 5.5, GPT-6, and Kimi K3.1 Pro at much lower prices
Source: The Decoder (opens in a new tab) · Jonathan Kemper
Intel Summary
Alibaba has launched Qwen4-Max, a flagship AI model featuring 2.4 trillion total parameters with 71 billion activated per token. The model uses a hybrid architecture designed to reduce long-context inference costs and ranks just behind Kimi K3.1 Pro on the Artificial Analysis Intelligence Index. Alibaba also announced plans to open-source smaller models in the Qwen4 family.
Why It Matters
The sparse activation approach and hybrid architecture intensify price-to-performance competition among frontier model providers. If smaller open-source Qwen4 variants match the efficiency gains of the flagship release, developers and enterprise builders will gain access to lower-cost frontier-class capabilities.
Organizations & Entities
Topics
Related Intelligence
- ReportAlso involving Alibaba
Chinese tech firms face a double hit as chip shipments stop and overseas cloud access disappears
The Decoder reports that intensified US export controls are curbing AI chip access for Chinese companies, leading TSMC and Samsung to halt shipments to Chinese customers. The restrictions contributed to Moonshot AI withdrawing from a $150 million deal in Indonesia, while a New York Times investigation found Alibaba routed Nvidia chips to Chinese military researchers via front companies.
The Decoder - ReportAlso involving Artificial Analysis
OpenAI launches GPT-6.2 Astra and Sol with lower API prices and some benchmark gains
OpenAI has released GPT-6.2 Astra and Sol, introducing two new models with lower API pricing and enhanced capabilities compared to previous versions. According to the independent Artificial Analysis Intelligence Index, Astra ranks second behind Claude Fable 5.2, and marks the first released model to reach OpenAI's internal 'High' threshold for AI self-improvement risk.
The Decoder - ReportAlso involving Alibaba
Alibaba CEO Warns of Impact From New U.S. Chip Rule
Alibaba Group CEO Eddie Wu stated that a new U.S. rule prohibiting Chinese companies from accessing advanced Nvidia AI chips via overseas data centers will affect Alibaba's upcoming products and services. In response, the company is accelerating efforts to reduce its reliance on Nvidia hardware.
The Information - ReportAlso involving Artificial Analysis
Free Chinese AI Can Hack Companies While Frontier Developers Face Diminishing Returns
The Information reports that freely downloadable Chinese AI models have developed sufficient capability to penetrate corporate networks, according to Ariel Zeitlin of automated penetration testing startup Tenzai. Meanwhile, benchmarking firm Artificial Analysis observes that frontier developers face rising costs and diminishing returns when improving top-tier cybersecurity capabilities, narrowing the lead of proprietary models.
The Information