metadata
license: cc-by-4.0
pretty_name: CPIBench-0 — Multimodal Agent and Model Rollouts
language:
- en
size_categories:
- n<1K
tags:
- multimodal
- image
- text
- tabular
- datasets
- agents
- agent-evaluation
- cyber-physical-systems
configs:
- config_name: default
data_files:
- split: train
path: data/sample.parquet
CPIBench-0
CPIBench-0 is our first benchmark for evaluating how frontier models and agent harnesses perform in operations that rely heavily on cyber-physical systems.
Built from real projects across mechatronics, manufacturing, materials, and energy, CPIBench-0 tests models and agent harnesses in multimodal workflows where cyber-physical signals, tools, and real-world consequences intersect.
CPIBench-0 measures performance across connected cyber-physical workflows, from making reliable operational decisions to progressing safely toward real-world action.
The public dataset repo contains a selection of replayable example eval tasks and rollouts. Please contact us to perform the full eval suite.