Summary
According to a post on X by @SciFi, Google Cloud AI Research has launched PLE (Physical-Engineering's Last Exam), a benchmark of 85 long, agentic tasks in electrical, civil, structural and mechanical engineering. The tasks were built with practising engineers through Mercor, and solving one means producing a deliverable a licensed reviewer would sign off on. The post reports that across 14 frontier models and seven agent harnesses, the best setup, Claude Opus 5.5 in Claude Code, averages only 50.6%, spending about 27M tokens. The snippet is cut off, so the full methodology is not shown.
AI-assisted summary based on the listed source.
What happened
PLE (Physical-Engineering's Last Exam) has 85 long, agentic tasks in electrical, civil, structural and mechanical engineering, built with practising engineers through Mercor, where solving a task means producing a deliverable a licensed reviewer would sign off on. Across 14 frontier models and seven agent...
Why it matters
The benchmark tests AI agents on engineering deliverables that a licensed professional would approve, a more demanding standard than short questions, which matters for engineering firms and the builders who work with them. The reported 50.6% top average suggests current agent setups still fall well short on these tasks. Watch for the full methodology and whether other labs or engineering teams publish results on PLE.
Signal Intelligence
Signal Strength 90%
Technical label WATCH
Public Interest 44
Category RESEARCH
Reader Depth TECHNICAL
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
Recognizable Entity Score 0
Practical Impact Score 54
Novelty Interest Score 94
Consequence Score 50
Curiosity Score 16
Shareability Score 55
Why this is here
VQV surfaced this signal because it is recent, relevant to AI Search, connected to Scobleizer News Pulse (X).
