Get all your news in one place.
100's of premium titles.
One app.
Start reading
Windows Central
Windows Central
Technology
Kevin Okemwa

"We made a mistake in not being more transparent": OpenAI secretly accessed benchmark data, raising questions about the AI model's supposedly "high scores" — after Sam Altman touted it as "very good"

OpenAI and ChatGPT.

In the next couple of weeks, OpenAI is slated to unveil o3 mini, the smaller version of its o3 series with advanced reasoning capabilities across math, science, and coding. CEO Sam Altman claims the model is "very good," potentially beating the performance of the ChatGPT maker's o1 series. The company also announced that it would launch the application programming interface (API) and ChatGPT for the model simultaneously.

While details about the AI model remain slim, reports suggest OpenAI secretly funded and accessed the FrontierMath benchmarking data, raising concerns about whether the company used the data to train o3 (via Search Engine Journal). The AI model received high scores across a wide range of benchmarks, but if the rising concerns are anything to go by, the shared results might be a fabrication of the model's reality.

Sign up to read this article
Read news from 100's of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
One subscription that gives you access to news from hundreds of sites
Already a member? Sign in here
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.