RWE-bench (Real World Engineering Bench)
by Xplore
What is RWE-bench (Real World Engineering Bench)? RWE-bench (Real World Engineering Bench) is an agents & tool use eval made by Xplore. Tests frontier models on 20 physical engineering design tasks: mechanical parts in CAD, high-speed PCB routing such as DDR memory interfaces, and aerospace CFD and heat-transfer simulations. Each model works in its vendor's own agent harness with CAD/CAE tools, and every submission is checked by a deterministic verifier that runs physics solvers where needed, with no LLM or human grading. Tasks, verifiers and reference solutions are not public; outside evaluation is by contacting Xplore.
Editor note
Tests frontier models on 20 physical engineering design tasks: mechanical parts in CAD, high-speed PCB routing such as DDR memory interfaces, and aerospace CFD and heat-transfer simulations. Each model works in its vendor's own agent harness with CAD/CAE tools, and every submission is checked by a deterministic verifier that runs physics solvers where needed, with no LLM or human grading. Tasks, verifiers and reference solutions are not public; outside evaluation is by contacting Xplore.
- What it tests
- Whether AI agents can do real mechanical, electrical and aerospace engineering work that passes physics checks.
- Original source
- xplore.ai/rwe-bench
- Last updated
- Launched 2026-10-27
- License
- Not checked yet — see source
Related agents & tool use evals
- GAIA — General assistant skills: browsing, file handling, multi-step reasoning.
- WebArena — Autonomous web navigation and task completion.
- OSWorld — Computer-use agents operating real operating systems and apps.
- Terminal-Bench — Agents working in the shell end to end.
Discussion
No one has discussed RWE-bench (Real World Engineering Bench) yet. Have you run it? Share what it caught, what it missed or how you set it up.