Thread by @OpenAIDevs on Thread Reader App
OpenAI announced GPT-Live-1, a voice API that lets AI agents listen and speak simultaneously for more natural conversations, plus rolled out developer features like customizable keyboard shortcuts, a skills system for Codex, and a multi-provider LLM spec called Open Responses.
- GPT-Live-1 handles simultaneous listening and speaking in a single model, reducing latency and letting users interrupt or add details mid-conversation without waiting for the AI to finish
- Codex now supports reusable skills (instruction bundles following the agentskills.io standard) that can be called directly or selected automatically based on user prompts
- Open Responses is an open-source spec for building multi-provider LLM interfaces, letting developers build agentic systems without rewriting code for each model
- New efficiency improvements include GPT-5-Codex-Mini for cost savings, 50% higher rate limits across ChatGPT tiers, and improved Git workflow integration in the editor
This is a collection of OpenAI product announcements pulled from Twitter threads spanning several months. The headline feature is GPT-Live-1, a voice API that handles listening and speaking simultaneously, which means you can interrupt it mid-response or add details without waiting for it to finish talking. It filters out background noise so conversations work in noisy environments like cafés. The model combines listening and speaking in one pass rather than shuttling between separate components, which cuts down latency and makes interactions feel more natural.
The other announcements cover developer tools. Codex—their coding-focused model—now supports "skills," which are reusable bundles of instructions and scripts that you can call directly or let the model choose automatically. They released GPT-5.2-Codex for the Responses API, positioned for complex tasks like refactoring and finding vulnerabilities. They also announced Open Responses, an open-source spec for building multi-provider LLM interfaces so you don't have to rewrite your stack for each model. On the usage side, they bumped rate limits 50% higher for ChatGPT Plus users and released a cheaper Mini version of their main Codex model.
The updates show OpenAI pushing toward making AI interactions feel less mechanical—whether through voice that doesn't require turn-taking or through developer tools that reduce friction when switching between models and providers. The skills system specifically tries to standardize how reusable components work across different tools.