TLDR AI · QUICK LINKS
WE FINALLY HAD A MOMENT TO RUN OUR SYSTEM WITH GPT-5.2 X-HIGH ON ARC-AGI-2!
Summary from the TLDR AI newsletter
Poetiq saw results as high as 75% at under $8/problem using GPT-5.2 X-High on the full PUBLIC-EVAL dataset.
TLDR AI · QUICK LINKS
Poetiq saw results as high as 75% at under $8/problem using GPT-5.2 X-High on the full PUBLIC-EVAL dataset.