GPT-6 Astra gets 3% on the FrontierMath Erdős Benchmark, while every other Model(that was tested) got 0%

View original

OpenAIModel release

benchmark

Time & source

Published: 09/04, 16:52
First discovered: 09/04, 20:00
Type: Dev community · RSS

AI summary

GPT-6 Astra achieved a 3% score on the FrontierMath Erdős Benchmark, significantly outperforming all other tested models, which scored 0%. This result highlights Astra's exceptional mathematical capabilities, although the overall low scores indicate that there is still substantial progress to be made in AI mathematics. The performance suggests a promising trajectory for AI development by the end of 2026.

https://preview.redd.it/dhi42fp49jnh1.png?width=666&format=png&auto=webp&s=1c67e49c0f60575ad70ab60d30a775e36462800c

So yeah, Astra is extremely good at mathematics. But we still have a long way to go. I wonder where we'll be at the end of 2026.