OpenAI claims GPT-6 Astra scores high on everyday tasks like shopping and meal planning
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…
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
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.
Model pages: GPT-6 Astra → · Muse → · Claude Duet →
The story so far
2 episodes →- OpenAI claims GPT-6 Astra scores high on everyday tasks like shopping and meal planningthis story
OpenAI says its AI is getting better at doing things humans do, like shopping and meal planning
cnn.com · 2 November 2026
Loading the full article…
This text was published by cnn.com and written by Lisa Eadicicco. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
1source- Reddit discussionreddit.com
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Deep Principle releases Hydrogen-1.0, a 753B open-weight science model · 1 src
- Google releases Gemini 4 Argon and announces Gemini 4 Carbon · 2 src
- Ggml-org adds Clover model support to llama.cpp · 1 src
- Google rolls out Genie 4 world model to trusted testers · 1 src
- Anthropic reportedly prioritizes safety over image generation for Claude · 1 src
Comments
via GitHub Discussions