“AI Disruption” Publication 10,500 Subscriptions 20% Discount Offer Link.
GPT-6 Astra is the most capable model to date. It delivers outstanding performance in computer use, browsing, software engineering, scientific research, and professional work. It is especially strong at executing multi-step workflows across code, the browser, and professional software, acting with care, respecting task boundaries, and communicating transparently.
When instructions leave room for interpretation, it uses context to fill in ordinary gaps and asks targeted questions when the answer could affect the outcome. It can incorporate new requirements, change direction when asked, and answer side questions without losing sight of the overall task.
For a detailed model introduction, see the previous article.
1. Use a lower reasoning effort
For ordinary tasks, default to reasoning.effort: "low".
