23
← Back to the timeline
Daily archive

AI news on February 23, 2026 · Monday

10 stories — deduplicated across sources, ranked by significance, every source cited.

Why we no longer evaluate SWE-bench Verified

Opinion·OpenAI Blog·Feb 23O

OpenAI announces Frontier Alliance Partners

Products·OpenAI Blog·Feb 23O

American AI Industry Trembles as Deepseek Prepares to Release New Model - Futurism

Models·DeepSeek News·Feb 23D

Exclusive: China's DeepSeek trained AI model on Nvidia's best chip despite US ban, official says - Reuters

Models·DeepSeek News +1·Feb 23DY

The persona selection model - Anthropic

Other·Anthropic News·Feb 23A

Alibaba AI User Surge: Why 58M Qwen Users Change the Outlook - TIKR.com

Models·Qwen News·Feb 23Q

Five new sovereign AI models signal India’s bold leap at AI Summit - Forbes India

Other·Sarvam AI News +1·Feb 23SS

Anthropic Education Report: The AI Fluency Index

Open Source·Anthropic·Feb 23A

Zhipu apologises for three GLM-5 missteps, unveils compensation plan

Other·The Standard (HK)·Feb 23T