ChatGPT agent System Card: OpenAI’s agentic model unites research, browser automation, and code tools with safeguards under the Preparedness Framework.
Introducing ChatGPT agent: it thinks and acts, using tools to complete tasks like research, bookings, and slideshows—all with your guidance.
We study how training on incorrect responses can cause broader misalignment in language models and identify an internal feature driving this behavior—one that can be reversed with minimal fine-tuning.
We are replacing the existing GPT-4o-based model for Operator with a version based on OpenAI o3. The API version will remain based on 4o.
Codex is a cloud-based coding agent. Codex is powered by codex-1, a version of OpenAI o3 optimized for software engineering. codex-1 was trained using reinforcement learning on real-world coding tasks in a variety of environments to generate code that closely mirrors human style and PR preferences, adheres precisely to instructions, and iteratively runs tests until passing results are achieved.
Vi introduserer Codex: en skybasert programvareutviklingsagent som kan jobbe på mange oppgaver parallelt, drevet av codex-1. Med Codex kan utviklere rulle ut flere agenter samtidig for å uavhengig håndtere kodeoppgaver som å skrive funksjoner, svare på spørsmål om kodebasen din, løse feil og foreslå pull-forespørsler for gjennomgang.
HealthBench is a new evaluation benchmark for AI in healthcare which evaluates models in realistic scenarios. Built with input from 250+ physicians, it aims to provide a shared standard for model performance and safety in health.
OpenAI o3 og o4-mini representerer et betydelig gjennombrudd i visuell persepsjon ved å resonnere med bilder i tankegangen.
OpenAI o3 og OpenAI o4-mini kombinerer toppmoderne resonnement med komplette verktøyfunksjoner – nettlesing, Python, bilde- og filanalyse, bildegenerering, canvas, automatiseringer, filsøk og minne.