techmeme.com

The Arc Prize Foundation says its new ARC-AGI-2 test stumps most AI models; humans get 60% of the questions right but GPT-4.5 and Claude 3.7 Sonnet score ~1% (Maxwell Zeff/TechCrunch)

Maxwell Zeff / TechCrunch: The Arc Prize Foundation says its new ARC-AGI-2 test stumps most AI models; humans get 60% of the questions right but GPT-4.5 and Claude 3.7 Sonnet score ~1%  —  The Arc Prize Foundation, a nonprofit co-founded by prominent AI researcher François Chollet, announced in a blog post …

Mar 25, 2025 - 20:40
 0
The Arc Prize Foundation says its new ARC-AGI-2 test stumps most AI models; humans get 60% of the questions right but GPT-4.5 and Claude 3.7 Sonnet score ~1% (Maxwell Zeff/TechCrunch)

Maxwell Zeff / TechCrunch:
The Arc Prize Foundation says its new ARC-AGI-2 test stumps most AI models; humans get 60% of the questions right but GPT-4.5 and Claude 3.7 Sonnet score ~1%  —  The Arc Prize Foundation, a nonprofit co-founded by prominent AI researcher François Chollet, announced in a blog post …

Read More

Tags:

Previous Article

OpenAI starts rolling out GPT-4o-powered "Images in ChatGPT" to all tiers includ...

Next Article

Meta rolls out a pilot program on Instagram designed to let US schools flag any ...

Related Posts

Cybersecurity startup Andesite raised an additional $23M seed, bringing its total funding to $38.25M, and unveils a security operations center product (David DiMolfetta/Nextgov/FCW)

Cybersecurity startup Andesite raised an additional $23...

Feb 12, 2025  0

Experts say Sean Cairncross' nomination as national cyber director signals the White House's ONCD will lead US cyber policy across the government (Suzanne Smalley/The Record)

Experts say Sean Cairncross' nomination as national cyb...

Mar 8, 2025  0

Google and UCB researchers detail "inference-time search", which some call a fourth AI scaling law, though experts are skeptical of its usefulness in many cases (Kyle Wiggers/TechCrunch)

Google and UCB researchers detail "inference-time searc...

Mar 20, 2025  0