# OpenAI claims GPT-6 Astra scores high on everyday tasks like shopping and meal planning

Digest AI · Generative AI & Models · published 2026-11-02T17:41:16Z

Canonical: https://digestai.news/story/openai-claims-gpt-6-astra-scores-high-on-everyday-tasks-like-shopping

## Summary

OpenAI announced on Monday that its GPT-6 Astra model outperformed earlier versions on new benchmarks for real-world tasks. The company claims GPT-6 Astra scored 93.1% on finding a kindergarten after a move, 98% on managing flood volunteer coordination, and 90.4% on creating a diabetes-friendly meal plan. These scores come from an internal test designed to measure how AI handles daily activities like shopping, budgeting, and form-filling, which OpenAI says are often left to humans.

The new benchmark, called **Everyday Tasks Evaluation**, was developed after surveying 4,000 ChatGPT users and 100 experts to identify tasks AI could assist with. OpenAI’s researcher Jenny Yang noted that GPT-5.2 scored just 49.3% on similar tasks, highlighting progress. The company also cited Meta’s Muse agent—downloaded 10 million times in the US in 40 days—as part of a competitive push to turn ChatGPT into a hands-off assistant. Safety concerns persist, however, as AI agents navigate complex, real-world scenarios.

## Key points

- GPT-6 Astra scored 93.1% on finding a kindergarten after a move, per OpenAI’s internal benchmark
- OpenAI surveyed 4,000 users to design tests for everyday tasks like shopping and meal planning
- Meta’s Muse agent hit 10 million US downloads in 40 days, per Sensor Tower data

## Why it matters

Higher scores on real-world tasks could accelerate AI adoption for mundane chores, but subjective benchmarks and safety risks remain unresolved.

## Sources

1. [OpenAI says its AI is getting better at doing things humans do, like shopping and meal planning](https://cnn.com/2026/11/02/tech/openai-chatgpt-personal-assistant-ai) (cnn.com, 2026-11-02)

Part of the developing story: [Rise of Practical AI Agents](https://digestai.news/thread/meta-releases-muse-agent-amazon-blocks-it) (2 stories)

## Cite

Digest AI, "OpenAI claims GPT-6 Astra scores high on everyday tasks like shopping and meal planning", 2 November 2026, https://digestai.news/story/openai-claims-gpt-6-astra-scores-high-on-everyday-tasks-like-shopping

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/openai-claims-gpt-6-astra-scores-high-on-everyday-tasks-like-shopping.json