← Front page

The Artificial Intelligence Show · Tuesday, September 8, 2026

AI Models Show Progress in Solving Unfamiliar Problems and Mathematical Reasoning

Recent advancements in AI models, exemplified by GPT-6 Astra, demonstrate a significant leap in problem-solving capabilities. Astra achieved a 99.9% score on the ArKI-3 benchmark, which tests an AI's ability to learn and strategize in unknown scenarios without explicit rules. Furthermore, it scored 97.6% on a challenging mathematics test, indicating improved reasoning for scientific discovery.

companyOpenAI

The tape

3 quotes
Notably, learning how to solve unfamiliar problems.
Astra scored a 99.9% on that under OpenAI's evaluation setup.
In mathematics, Astra scored 97.6% on Frontier Math's hardest tier. That's up from 83%.
Heard on The Artificial Intelligence Show — “#237: GPT-6 Astra, NYC Bans AI in Schools, Trump Goes All-In on Data Centers & Lawyers Under Pressure to Pass on AI Savings, published Tuesday, September 8, 2026. Heardvine summarizes and quotes with attribution and timestamps, and links to the original everywhere.
Transcribed via Gemini audio transcription · $0.07