BriefsNewsAvailable now
Google announces Physical-Engineering's Last Exam benchmark for AI agents
Google introduced Physical-Engineering's Last Exam, a benchmark and open competition to test AI agent performance in mechanical, electrical, civil, and structural engineering.
Summary
Google Cloud announced Physical-Engineering's Last Exam (PLE), a benchmark and open competition developed with Mercor to test frontier AI models on mechanical, electrical, civil, and structural engineering tasks. Google released 29 public tasks, grading tools, and an open competition leaderboard. Baseline tests show all frontier models score below 50% on the private test set.
What changed
- Google introduced Physical-Engineering's Last Exam, a new benchmark and open competition for AI agents.
- Google released 29 public tasks, grading tools, and an open competition leaderboard at ple-bench.com.
Availability
Available now“Download the 29 public tasks, grading scripts, and evaluation container from our GitHub repository .”
Founder takeaway
There is nothing to build or buy yet, but teams developing AI agents for engineering can test their setups against the new benchmark.
Source
Researched and drafted with AI from the official sources above, then checked automatically before publishing. Spotted a mistake? Tell us.
Related
More briefs
OpenAI launches an internal AI security sprint and expands defense access
OpenAI announced a security sprint with outside partners, increased bug bounty payouts, and expanded AI access for defenders.
OpenAI1 min read
Angle Health uses ElevenLabs AI agents to resolve 90 percent of provider calls
Angle Health ran a pilot where ElevenLabs AI voice agents handled routine provider calls without human transfers.
ElevenLabs1 min read
Nvidia and partners expand physical and agentic AI services for cities
Global cities are using Nvidia physical and agentic AI tools for municipal services, digital twins, and traffic management.
NVIDIA1 min read
One email, every Monday
Briefs are published here through the day. Monday Heist is the email: one useful AI workflow, the files to use it, and the catch. The first issue is being prepared; subscribers get it first. Free, and you can leave in one click.
One email. Every Monday.