InfoNewsLLM-specific
Alibaba touts 62% lower task costs with Qwen4-Max, adds cyber safeguards
- Published
- Record updated
Summary
Alibaba released Qwen4-Max, a 2.4-trillion-parameter mixture-of-experts model, and claims it cuts the average cost per professional task by 62%, from $4.43 to $1.67, against its predecessor Qwen3.8-Max. The company says it filters training data to keep defensive cybersecurity content and remove offensive content, and that code security checks raised the Last Secure Code Bench pass rate from 22% to 31.8%. Analysts cited in the article call the cost figure a vendor benchmark and say the security separation needs independent validation.
Related items
- InfoAltimeter's Brad Gerstner aims to invest $5 billion in Anthropic IPOSame vendor · CNBC Technology
- InfoOpenAI inks agreements with four independent evaluators to boost company's safety workSame vendor · CNBC Technology
- InfoClaude Code with Fable 5.2 scores 36.9% SecPass and 64.8% FuncPass on our coding tasks | Endor LabsSame vendor · Endor Labs Blog
- InfoAnthropic’s Wall Street whisperer: Meet CFO Krishna RaoSame vendor · CNBC Technology
- MediumCursor Supply Chain Attack Started at Fireworks AISame vendor · SecurityWeek