Physical AI
Analysis
An independent AI evaluation company

Independent analysis
of physical AI.

How well can AI design physical systems
and operate them? We evaluate the capabilities
that connect intelligence to the physical world.

Explore the evidence
Robotics · CAD · PCB design · SimulationCapability, tested.

01 / Research focus

From generated designs
to observable outcomes.

We study AI where decisions meet physical constraints. Robotics is our flagship focus, alongside the tools that design and simulate engineered systems.

01.1Flagship focus

Robotics

Planning, control and interaction: can an AI system turn an instruction into successful behavior?

Current published work includes simulation. Real-world execution is a separate question.

01.2Research scope

CAD & engineering

Geometry, assemblies and engineering intent: can AI produce designs that hold up beyond a convincing image?

Design artifacts and simulation are evaluated within their stated scope.

01.3Research scope

PCB design

Circuits, components and layout: how well can AI work within the constraints of electronic design?

A generated board design is distinct from fabricated and tested hardware.

02 / Selected published work

Evidence you
can inspect.

Published evaluations from RemakeBench, our benchmark for AI building and remaking games, with work extending into physical design and simulation.

Rendered simulation of a robot pick-and-place attempt, GPT-6.1 Sol.
Robot pick-and-place · GPT-6.1 Sol · Rendered simulation
Model comparison

Opus 5.5 vs GPT-6 Astra vs Sonnet 5.5 vs GPT-6.1 Sol

A published comparison across games, visual scenes, CAD, PCB and simulated robot tasks. Each round preserves its own evidence and outcome.

Read the evaluation

Simulation is distinct from physical execution. Task-specific outcomes; no aggregate score.

Claude Opus 5 generated storm-city scene.
Storm-city scene · Claude Opus 5 · Generated graphics
Workflow evaluation

Claude Opus 5 across nine game-building tests

Nine frozen game-building and graphics tasks, with recorded attempts, configuration disclosures and limitations.

Read the evaluation

A workflow comparison with disclosed tools. Single recorded attempts do not establish reliability.

Every result belongs to a particular task, setup and scope.

Visit RemakeBench

03 / Our approach

Make the test clear.
Make the evidence
public.

Useful evaluation starts with a precise question and ends with something others can inspect. These principles guide our work.

  1. 01

    Consistent conditions

    Compare systems on the same tasks, with the same constraints and clearly stated evaluation criteria.

  2. 02

    Disclosed configurations

    Describe the model, tools, prompts and environment needed to understand what was actually tested.

  3. 03

    Observable outcomes

    Inspect the resulting artifact or behavior. Distinguish design quality, simulated performance and physical execution.

  4. 04

    Failures & public evidence

    Show the limitations alongside the successes, and link the evidence that supports each conclusion.

04 / Our benchmarks

One company.
Distinct questions.

Physical AI Analysis is the parent company. Our benchmark brands investigate specific capabilities.

AI building games

RemakeBench

Evaluating AI at building and remaking games.

Published work
AI playing games

GameplayBench

A benchmark brand focused on AI playing games.

05 / Get in touch

Better questions.
Better evaluations.

Research questions, evaluation proposals
or a result worth looking into.

remakebench@gmail.com