GPT‑6 Astra announced: why it is more than a stronger chatbot
OpenAI’s new model advances professional work, computer use, coding and science in one system. The important shift is task completion, not just answer quality.
Representative editorial image · Not a real team photoOpenAI says GPT‑6 Astra is rolling out from a limited set of organisations to ChatGPT Plus, Pro, Business and Enterprise, as well as developers through the API and enterprise cloud partners. The API model is gpt-6-astra, priced at $10 per million input tokens and $50 per million output tokens.
The change is broader than better prose: Astra is positioned for computer interaction, terminal and coding work, long-context retrieval and scientific problem solving. OpenAI reports 72.6 on OSWorld, 57.9 on Terminal-Bench, 97.6 on FrontierMath Tier 4 and 99.9 on ARC-AGI-3. These are provider-published results; independent evaluation, latency, cost and failure behaviour still matter.
For a venture, the useful question is not which model is strongest, but which part of a workflow can be delegated, where human approval remains, how errors are reversed and what a completed task costs. OpenAI also says Astra reaches its Critical cyber capability threshold, so access control, logs, data boundaries and evaluation are part of the product—not an afterthought.