Live Feeds
● LIVE Updated 1d ago · 12 sources tracked

China's Zhipu is closing in on top U.S. AI models with Anthropic and OpenAI held back

China's Zhipu AI has released the GLM-5.2 model, which competes with top U.S. systems in specific tasks. The model is reportedly on par with Anthropic's Claude Mythos regarding cybersecurity and vulnerability detection. This progress occurs as Trump administration policies restrict U.S. AI models.

RSS Source map (10)

What changed

Zhipu AI released GLM-5.2, matching some performance levels of Anthropic's restricted models.

Live updates

  1. Zhipu AI Closes Gap With U.S. Models in Cybersecurity

    China's Zhipu AI has released the GLM-5.2 model, which competes with top U.S. systems in specific tasks. The model is reportedly on par with Anthropic's Claude Mythos regarding cybersecurity and vulnerability detection. This progress occurs as Trump administration policies restrict U.S. AI models.

    What's confirmed:

    • Zhipu AI developed the GLM-5.2 model.
    • GLM-5.2 is an open-weight model.
    • Chinese AI models have reached parity with Anthropic in cybersecurity tasks.

    Still unconfirmed:

    • GLM-5.2 falls short of Anthropic and OpenAI models in tasks other than cybersecurity.
    confidence 85%