OpenAI API Update Introduces GPT-5.1 as Default Reasoning Engine, New Codex Variants, and Assistants API Sunset Plan
GPT-5.1 powers logic-heavy workflows, new Codex variants for agents
OpenAI has rolled out a significant API update that establishes GPT-5.1 as the new default reasoning engine, replacing GPT-5 for logic-heavy and multi-step workflows. This shift aims to enhance performance in tasks requiring deep reasoning and complex decision-making. The update also introduces new Codex variants, including GPT-5.4 mini and nano models, specifically designed for agentic and high-volume AI development. These models are optimized for efficiency and scalability, catering to developers building autonomous agents and large-scale applications.
In addition to model updates, OpenAI announced a plan to sunset the Assistants API, consolidating all its functionality into the Responses API. This move provides developers with a clear migration path rather than a hard cutoff, ensuring continuity and reducing disruption. The consolidation simplifies the API landscape, offering a unified interface for building AI-powered assistants and agents. Developers are encouraged to transition to the Responses API to leverage improved features and long-term support.
- GPT-5.1 replaces GPT-5 as the default reasoning engine for logic-heavy and multi-step workflows.
- New GPT-5.4 mini and nano Codex variants target agentic and high-volume AI development.
- Assistants API to be sunset, with all functionality migrating to the Responses API.
Why It Matters
Unified API and optimized models streamline AI development, enabling more efficient agentic and high-volume workflows.