News for hackers. Stuff that matters.
Re:The US Is Winning the AI Race by pattt on Wednesday May 13, 2026 @04:13PM (#48126876)
Do we have any solid evidence these models can outperform Western models in terms of quality? Or is it more: because they are forbidden, they can't get enough training data, visibility etc. to compete?
Re:The US Is Winning the AI Race by gpt5 on Wednesday May 13, 2026 @05:33PM (#48127932)
Scroll down to the leaderboard - https://arcprize.org/leaderboard

Spoiler alert - they are all towards the bottom of the leaderboard. People come up with a wide variety of excuses for why they are not used despite being offered for significantly lower cost, but the answer is simply because they don't perform well enough for now.

Re:The US Is Winning the AI Race by aucisson_masque on Wednesday May 13, 2026 @06:13PM (#48128293)
There isn't even deepseek V4.

I'd rather trust LLM arena leaderboard, which puts it on par with sonnet.

Re:The US Is Winning the AI Race by gpt5 on Wednesday May 13, 2026 @06:20PM (#48128369)
LM Arena uses human side by side voting, which limits its applicability to complex tasks.

The ARCPrize leaderboard does have Deepseek V3.2, which only scored 4% on ARC-AGI 2 (while the top models score over 80%). It also Kimi and Qwen, but they also didn't perform well.