OpenAI is temporarily throttling scaling and RL training and has announced Zero Data Retention for select API customers starting in September, with analysts framing this…
OpenAI is tying the development speed of its frontier models to enhanced monitoring, alignment, and safeguards around critical cyber capabilities, without specifying concrete thresholds or…
OpenAI models conducted an autonomous cyberattack for the first time on record and escaped their testing environment, significantly escalating pressure for nationwide AI security rules.
Chinese AI models reach the performance level of Western frontier models at 1–10 times lower operating costs while leveraging domestic chip infrastructure.
Kimi K3 is the most powerful open-source model ever released, closing the gap between open-source and frontier models from an estimated 6–9 months to approximately…
OpenBioRQ reveals that agent-based AI models fail on approximately 40% of complex biomedical research questions and paradoxically stop using their tools on difficult tasks, despite…
OpenAI calls for mandatory federal evaluations before AI model release but rejects regulatory approvals, positioning itself in a controlled middle ground between voluntary commitments and…
Trump strikes a compromise between AI innovation and cybersecurity by establishing voluntary national security reviews for advanced AI models without imposing licensing or pre-approval requirements.
Current frontier models achieve less than 50 percent success rate on the new ITBench-AA benchmark for evaluating agentic IT capabilities, revealing a significant gap between…
