OpenAI released dozens of updates and new models for API developers
OpenAI has announced a significant volume of updates and releases for its developers over the past six months. During this period, over 30 new models, features, and improvements for the API have been released. The main focus has been on expanding capabilities for building complex applications and agents. Key highlights include several new versions of GPT models, including GPT-5.5, GPT-5.4 mini, and GPT-5.4 nano, as well as specialized models for real-time operation (GPT-Realtime-2, GPT-Realtime-Whisper, GPT-Realtime-Translate). Furthermore, OpenAI has significantly improved agent development tools by introducing the Agents SDK harness & sandbox, and adding Skills capabilities to the Responses API. These updates are aimed at enabling developers to create more complex, multifunctional, and integrated solutions.
Why it matters
- —Показывает высокую скорость инноваций OpenAI, что важно для разработчиков.
- —Новые модели (GPT-5.5, nano) расширяют возможности для нишевых и ресурсосберегающих задач.
- —Улучшения для создания агентов (Agents SDK) делают разработку сложных ИИ-систем более доступной.
Key facts
- Выпущено более 30 моделей, функций и инструментов для API за последние 6 месяцев.
- Новые модели включают GPT-5.5, GPT-5.4 mini и GPT-5.4 nano.
- Добавлены инструменты для создания агентов: Agents SDK harness & sandbox.
- Улучшены возможности взаимодействия через Responses API и WebSocket mode.
The full text is in the original source. Here we provide a brief summary and key facts.
