Claims

What was said, typed and attributed. Statements as reported by the cited publishers — not verified by DecisionGateway.

Follow via RSS — the 30 newest claims in any reader, no account needed.

Claims by
statementOpenAI told investors that its adjusted operating margin rose from -150% to -109% from Q2 to Q3.OpenAIstatementOpenAI told investors its adjusted operating margin improved from -150% to -109% from the second quarter to the third quarter.OpenAIstatementThere were about 1,000 medium-severity alerts.OpenAIstatementSuch high-risk misaligned behavior has only been observed in synthetic evaluations.OpenAIstatementThe system triggered zero highest-severity alerts, which are designed to catch high-risk misalignment such as coherent scheming.OpenAIstatementOver the past five months, the system monitored tens of millions of internal agent trajectories.OpenAIstatementThe system automatically alerts on anomalous behavior that may not match user intent or violate internal safety policies, so that other safeguards can be quickly investigated and improved.OpenAIstatementThe monitoring system is powered by GPT-5.4 and can review an agent's chain of thought and actions within 30 minutes after an interaction completes.OpenAIstatementOpenAI has built and optimized a low-latency internal monitoring system to supervise its internally deployed coding agents.OpenAIstatementThere were about 1,000 medium-severity alerts.OpenAIstatementSuch high-risk misalignment behavior was observed only in synthetic evaluations.OpenAIstatementThe system triggered zero highest-severity alerts, which are designed to catch high-risk misalignment such as coherent scheming.OpenAIstatementOver the past five months, the system monitored tens of millions of internal agent trajectories.OpenAIstatementThe system automatically alerts on anomalous behavior that may not match user intent or that violates internal safety policies, to enable rapid investigation and improvement of other safeguards.OpenAIstatementThe system can review an agent's chain of thought and actions within 30 minutes after an interaction completes.OpenAIstatementThe monitoring system is powered by GPT-5.4.OpenAIstatementOpenAI has built and optimized a low-latency internal monitoring system to supervise its internally deployed coding agents.OpenAIstatementChatGPT and the API now generate more than 4 billion images per week, up from 3 billion when Images 2.5 launched in September.OpenAIstatementThe two new models increase image generation speed by 50%.OpenAIstatementOpenAI released ChatGPT Images 2.6 on November 4 and began rolling it out to ChatGPT and Codex users across web, desktop and mobile.OpenAI

20 of 21