OpenAI’s new research explains why language models hallucinate. The findings show how improved evaluations can enhance AI reliability, honesty, and safety.
OpenAI researchers are testing “confessions,” a method that trains models to admit when they make mistakes or act undesirably, helping improve AI honesty, transparency, and trust in model outputs.
OpenAI introduces CoT-Control and finds reasoning models struggle to control their chains of thought, reinforcing monitorability as an AI safety safeguard.
OpenAI introduces IndQA, a new benchmark for evaluating AI systems in Indian languages. Built with domain experts, IndQA tests cultural understanding and reasoning across 12 languages and 10 knowledge areas.
Learn how OpenAI’s Model Spec serves as a public framework for model behavior, balancing safety, user freedom, and accountability as AI systems advance.
In this post we’ll outline new OpenAI research in which agents develop their own language.
Discover how OpenAI's new safe-completions approach in GPT-5 improves both safety and helpfulness in AI responses—moving beyond hard refusals to nuanced, output-centric safety training for handling dual-use prompts.
We describe our latest thinking in the hope of helping other AI developers address safety and misuse of deployed models.
A new analysis from OpenAI reveals issues in SWE-Bench Pro, a popular coding benchmark, raising concerns about reliability and accuracy in evaluating AI models.
arXiv:2606. 12900v1 Announce Type: new Abstract: Large language models (LLMs) often hallucinate by generating factually incorrect or unfaithful content, posing significant risks to their safe use.
arXiv:2607. 04034v1 Announce Type: cross Abstract: The language models that underpin agents have seen a rapid rise in performance on function calling benchmarks.
OpenAI and the Lenfest Institute AI Collaborative and Fellowship program
OpenAI shares lessons from deploying long-running AI models, highlighting new safety risks, observed failures, and improved safeguards through iterative deployment.