NewSunSEO

xAI publicly released Grok 4.6, saying it matches GPT-5.6 Sol on several agentic coding and knowledge work benchmarks. The rollout lands first in Cursor.

Alibaba moved Qwen3.8-Max into general availability with a 2.4T MoE architecture, real benchmarks, and $2/$6 pricing, but the promised open-weight release is still missing.

A new 6,500-word essay from Mark Zuckerberg on personal superintelligence ignores real AI harms, and the public's low trust in tech executives makes that silence consequential.

Zhipu AI has released GLM-5.3, an open-weights coding model built on the GLM-5.2 base, with gains from post-training and emergent cybersecurity capabilities.