According to an OpenAI post, Perplexity has adopted GPT-6 Astra for end-to-end operational work. The model is now used to write communications, modify software, and monitor production systems, marking a shift toward broader autonomy in real-world deployments.
The post says Perplexity checks in with the model much less frequently than it did with earlier models. That suggests growing confidence in Astra's ability to handle multi-step tasks without constant human supervision.
Because the only source is OpenAI's own announcement, the article offers no independent data or differing perspective. Still, it signals that frontier models are being trusted with increasingly consequential system-level responsibilities.