Alibaba launches Qwen3.8-Max, signaling open-weight frontier shift

Alibaba has unveiled Qwen3.8-Max, a 2.4 trillion parameter model with advanced autonomous coding and multimodal capabilities, and plans to release open weights for it and a smaller version soon. Early evaluations place Qwen3.8-Max competitively, particularly in coding and vision tasks, though operational demands for large open-weight models remain significant. The release highlights a trend of Chinese AI labs, including Kimi, DeepSeek, and GLM, leading in the development of open-weight frontier models, while US labs maintain an edge in some closed offerings. AI

IMPACT Sets a new benchmark for open-weight models, potentially accelerating enterprise adoption of advanced autonomous coding capabilities.

RANK_REASON Frontier-lab model release with system card and early eval data. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on Smol AINews →

• Alibaba
• Claude Opus 4.7
• Claude Opus 5 Max
• DeepSeek V4 Flash
• Kimi K3
• Kimi K3 Max
• MiniMax
• Qwen
• Qwen 3.7 Max
• Qwen3.8-Max

• model release
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →