Alibaba Drops Another China AI Model With Breakthrough Performance

Alibaba Group Holding Ltd. released Qwen3.8-Max, a flagship AI model with 2.4 trillion parameters that ranks higher on some benchmarks than Moonshot's Kimi K3, and plans to release its weights for public download next week. The model excels in autonomous coding and long-horizon execution, activating only segments of its parameters to reduce costs, and trails only Anthropic's releases on Arena Leaderboard benchmarks.

Bloomberg -- Alibaba Group Holding Ltd. released its latest flagship AI model, with performance claims putting it alongside global leaders like Anthropic's Fable. Most Read from Bloomberg - Singapore's Gated Island for the Rich Is Marred by Decayed Homes https://www.bloomberg.com/news/features/2026-07-30/singapore-s-sentosa-cove-island-slides-from-monte-carlo-dream-to-decaying-homes?utm campaign=bn&utm medium=distro&utm source=yahooUS - India's Family Offices Embrace Profit-Sharing to Attract Top Talent https://www.bloomberg.com/news/articles/2026-07-31/india-billionaire-family-offices-offer-profit-sharing-to-attract-top-talent?utm campaign=bn&utm medium=distro&utm source=yahooUS - Crushed by Kospi Rout, Angry Koreans Rip Lee and Vow Not to Buy https://www.bloomberg.com/news/articles/2026-08-02/crushed-by-kospi-rout-angry-koreans-rip-lee-and-vow-not-to-buy?utm campaign=bn&utm medium=distro&utm source=yahooUS The new model, Qwen3.8-Max, is built around 2.4 trillion parameters and ranks higher on some benchmarks than the headline-grabbing Kimi K3 from Moonshot that was recently unveiled. Alibaba intends to release the weights for public download next week, marking the next major move in the intensifying race among China's AI developers that include Moonshot, DeepSeek and ByteDance Ltd. Alibaba's debut comes after Moonshot's Kimi K3 sent ripples through stock markets and Silicon Valley last month as Chinese models showed the capability of quickly catching up with the top models crafted by Anthropic and OpenAI, despite relatively constrained computing resources. DeepSeek also has just expanded access to its latest model, V4 Flash, last week. The new Qwen3.8-Max performs well in autonomous coding and long-horizon execution, and was able to independently execute a software engineering project over 16 days in internal testing, Alibaba said. The new system has also improved in efficiency, activating only segments of its 2.4 trillion parameters when in use to reduce computational costs and latency. Alibaba showed it ranking only behind Anthropic's releases in a pair of Arena Leaderboard benchmarks. Most Read from Bloomberg Businessweek - Trump's Arctic Mining Deal Signals a New Era of State Capitalism https://www.bloomberg.com/news/features/2026-07-31/alaska-road-permit-shows-how-trump-upends-business-norms?utm campaign=bw&utm medium=distro&utm source=yahooUS - How a Few Hundred Dollars Could Manipulate Election Prediction Markets https://www.bloomberg.com/news/features/2026-07-29/manipulation-threat-taints-kalshi-and-polymarket-claims-as-political-forecasters?utm campaign=bw&utm medium=distro&utm source=yahooUS ©2026 Bloomberg L.P.