… Hallucinations have been an ongoing problem for AI models, but OpenAI says its new GPT-5.5 Instant model has “significant improvements in factuality across the board.” The company claims that, based on “internal evaluations,” GPT-5.5 Instant produced “52.5% fewer hallucinated claims” than its Insta… …
… Like Glasswing, Daybreak isn’t built on just one AI model — OpenAI says “Daybreak brings together the most capable OpenAI models, Codex, and our security partners.” Daybreak also involves specialized cyber models, including GPT-5.5 with Trusted Access for Cyber and GPT-5.5-Cyber, which began rollin… …
… All this will be underpinned by a new model series: Gemini 3.5, whose initial entry Gemini 3.5 Flash should be available next month. The model is supposed to have significantly better coding capabilities than Gemini 3, which was released to great fanfare last November. …
… Separately, XBOW released data suggesting “frontier models have taken a major step forward in vulnerability discovery.” Meanwhile, Microsoft said its multi-model agentic setup, MDASH , was used to discover 16 CVEs in this week’s Patch Tuesday updates and is the leader on the CyberGym security evalu… …
… But the OpenAI play — and the pivot to AI generally — suggests a specific vision: The key place to make money isn’t training foundation models. The real money is inference, or using AI models to output results on data that isn’t in the training set. …
… When you say “think,” is it the models thinking or is it the people making the models thinking? The people designing those models. How are they building out, what is the veracity of the data source that they use to train the models? …
… OpenAI claims otherwise, as you might imagine. …