Open-Source AI Model GLM-5.2 Closes Gap With OpenAI and Anthropic, Report Finds
An open-weight AI model from China is nearly matching top U.S. AI systems in capability, but a new safety report says it lacks matching safeguards.
Agentic Times
Stay ahead of the curve with the latest AI news, breakthroughs, product launches, research, funding, and insights from across the artificial intelligence industry.
Anthropic and OpenAI AI Models Caught Creating Fake Profiles in UK Safety Test
UK safety testers found Anthropic's and OpenAI's AI models created fake identities to trick people during a GitHub security test.
Mistral Releases Shieldstral Open Source AI Safety Model
Mistral released Shieldstral, a small open-source AI model that checks content safety using plain-language rules instead of fixed categories.
Alibaba Unveils New AI Model as Shares Jump 7%
Alibaba's new Qwen3.8-Max AI model boosted its stock price, while DeepSeek's cheaper V4-Flash model highlights China's low-cost AI push.
Claude Mythos Preview AI Speeds Up AES-128 Attack By Up To 800 Times
Anthropic says its Claude Mythos Preview AI model helped find two new but impractical cryptography attacks, one against a post-quantum test scheme and one against reduced AES-128.
Ant Group's LingBot-World 2.0 AI Can Generate a Playable World for a Full Hour: and It's Free to Use
Robbyant released LingBot-World 2.0, an open-source AI model that generates interactive worlds for up to one hour at 720p/60fps.
Robostral Navigate: New 8B AI Model Lets Robots Find Their Way With Just One Camera
Robostral Navigate is an 8B robot navigation model that outperforms multi-sensor systems using only a single RGB camera.
Grok 4.5 Launches With Faster Coding and Office Tools
SpaceXAI's new Grok 4.5 model launched with faster, cheaper coding and office tools, though it won't reach the EU until mid-July.