Ce que OpenAI a publié en 2025

Gpt-image-1.5 Prompting Guide

The content provides a cookbook guide for prompting the GPT-image-1.5 model to achieve reliable image generation results. It focuses on techniques and best practices for using the model effectively. T…

Pioneering research on the path to AGI

OpenAI's research page outlines their mission to develop artificial general intelligence (AGI) that benefits humanity. They focus on deep learning to train AI systems for various tasks. The page highl…

GPT-5.2 Prompting Guide

The content provides a cookbook guide for prompting GPT-5.2 to achieve accurate and concise enterprise workflows. It focuses on techniques and examples to effectively use the GPT-5.2 model for busines…

Transcribing User Audio with a Separate Realtime Request

The content describes a cookbook for transcribing user audio using out-of-band Realtime sessions. This guide provides instructions on how to implement real-time transcription, which can be useful for …

Modernizing your Codebase with Codex

The OpenAI Codex CLI is now available to help developers modernize legacy codebases. This tool is designed to assist in updating and improving older code, making it more efficient and maintainable. Th…

Sora 2 is here

OpenAI has released Sora 2, a significant upgrade to its video and audio generation model. Sora 2 is more physically accurate, realistic, and controllable than its predecessor, with synchronized dialo…

GPT-5.4 mini Model | OpenAI API

GPT-5.4 mini introduces a faster, more efficient version of GPT-5.4 optimized for high-volume workloads, expanding the model’s applicability to resource-intensive tasks. The update includes a 10% pric…

GPT-5.4 nano Model | OpenAI API

OpenAI has introduced GPT-5.4 nano, a specialized model optimized for speed and cost efficiency in tasks such as classification, data extraction, ranking, and sub-agents. The model targets use cases w…

Models

OpenAI introduced gpt-5.4 as its flagship model for complex reasoning and coding, while also launching smaller variants (gpt-5.4-mini and gpt-5.4-nano) optimized for latency and cost efficiency. All l…

Build leading AI products on OpenAI’s platform

OpenAI has introduced GPT-5, a new generation of their AI models, including GPT-5.2, GPT-5.2 pro, and GPT-5 mini. This launch is significant as it offers advanced intelligence and multimodal capabilit…

GPT-5.4 Model | OpenAI API

OpenAI has introduced GPT-5.4, a frontier model designed for complex professional work. The model supports various reasoning efforts, including none (default), low, medium, high, and xhigh. For models…

GPT-5.2 Model | OpenAI API

OpenAI has introduced GPT-5.2 as their flagship model designed for coding and agentic tasks across various industries. This model supports different levels of reasoning effort, including none (default…

GPT-5.2 pro Model | OpenAI API

OpenAI has introduced GPT-5.2 Pro, a model designed to handle complex problems and support multi-turn interactions through the Responses API. This model is optimized for advanced API features and may …

GPT‑5: Amgen

GPT‑5: Creative Writing

GPT‑5: Coding & Design

GPT‑5: Medical Research

GPT‑5: Cursor

GPT‑5: First Look

Navigating health questions with ChatGPT

OpenAI has announced a new initiative to enhance ChatGPT's ability to navigate health-related questions, though specific details are not provided in the content. The update appears to focus on improvi…