Skip to content
Neural Daily

ResearchNews

OpenAI LifeVal benchmarks AI on everyday life tasks

🇭🇺 Olvasd el magyarul →

OpenAI's new benchmark measures how well AI helps people get everyday tasks done, from finding childcare to managing finances and government paperwork.

By: Mikola Máté Patrik · · 2 min read

OpenAI LifeVal cover image with the LifeVal wordmark, a crayon, a shopping cart, a bar chart and a map
Image: OpenAI · Press kit · Official cover image of the OpenAI LifeVal announcement · Source

What is LifeVal?

OpenAI introduced LifeVal on Monday (OpenAI). The benchmark measures how well AI can help with everyday tasks such as finding childcare, accessing government services, creating a budget, ordering groceries or designing a workout plan.

According to the company, a truly helpful personal assistant needs to do more than give advice. ChatGPT should be able to take actions that actually get things done. OpenAI says it used this evaluation over the past year to track and drive improvement in model performance on everyday tasks.

Where do the tasks come from?

The starting point was two large time use studies: the American Time Use Survey and the Multinational Time Use Study. OpenAI expanded on them with a new demographically representative survey of 3,995 respondents, asking where people would most want help from a personal assistant.

The resulting tasks span caregiving, health, education, household administration, personal finances, shopping and accessing government services.

What results does OpenAI show?

The announcement shows how current models perform on a few concrete workflows:

  • Enrolling in the right kindergarten after moving: 93.1%
  • Managing volunteers ahead of a flood: 98.0%
  • Finding a family home in Luzerne County that fits budget and commute: 100%
  • Setting up a remote game watch with grandma: 96.3%
  • A care and meal plan for a parent with a new diabetes diagnosis: 90.4%

The best recorded GPT-6 Astra score is 100% on several tasks, including booking a car inspection before a road trip, ordering halal groceries within budget with an itemized spreadsheet, and unsubscribing from email lists. The announcement does not give an overall LifeVal score.

What should you take from this?

OpenAI says it trained its models over the past year to become faster, more capable and more aligned on browser use, computer use and connector tasks. The goal is to move ChatGPT beyond a chat tool towards an agentic personal assistant.

Keep in mind that LifeVal is OpenAI's own evaluation, and the announcement does not compare results with other vendors' models. It is a useful signal of direction, not an independent ranking.

    Produced with the help of AI, reviewed by an editor. How we make our articles

    Reviewed and edited by the Neural Daily desk. About

    Related articles