GPT-5.4 holds the FrontierMath record. But for most math tasks, you don't need the most expensive model.
The US Department of Commerce has secured pre-release access to AI models from five major labs, including Google Deepmind and Microsoft, to test for national security risks before they become publicly available. This move is part of a broader effort to stay ahead of the rapidly evolving AI landscape and mitigate potential threats to national security.
Meta is introducing an AI-driven system to identify and protect minors on its platforms, using visual cues like body size and bone structure to estimate age. This move is part of a broader effort to enhance youth safety features and comply with regulatory pressures.
SAP is bolstering its data platform capabilities with the acquisition of Dremio and a significant investment in Prior Labs, signaling a major push into the AI-ready data platform market. This move is part of a broader effort to catch up with competitors in the AI space, with SAP committing over $1.1 billion to its AI initiatives.
OpenAI has rolled out a significant update to its ChatGPT model, replacing the default GPT-5.3 Instant with GPT-5.5 Instant, which boasts a 52.5% reduction in hallucinations on high-risk topics and improved benchmark scores. This update is set to significantly impact the accuracy and reliability of AI-generated responses, particularly in sensitive areas like medicine, law, and finance.
The building blocks for AI systems to train their own successors without human intervention are nearly in place, with a 60% chance of achievement by the end of 2028, and a 30% chance by 2027. This development could revolutionize the field of AI research, but also poses significant risks and challenges for human oversight and control.
Context window, intelligence, and reasoning depth matter most. Here are the models built for deep work.
Intelligence scores predict writing quality better than you'd think. Here are the models that produce the most natural, engaging content.