# Physical AI Design Index (PADI)

Canonical: https://codeatoms.ai/padi/
Publisher: CodeNinja Atoms (https://codeatoms.ai)

One question about system design: given the same short requirement for physical AI in a real operation, does Praxis produce a better system design than frontier AI on its own? PADI scores both on fresh tasks across 13 industries in three countries.

PADI is the benchmark CodeNinja uses to improve Praxis, one RSI loop (recursive self improvement loop) at a time. Scores are far from saturated: the goal is every check on every task in all 13 industries.

Latest RSI Loop 16 (2026-10-07): Praxis platform 36.5%, frontier AI 26.0% of checks passed.
All 5 RSI loops: Praxis platform 37.7%, frontier AI 34.6% over 546 checks per arm.
Industries measured: 10 of 13.

## Industries

| Industry | Praxis platform | Frontier AI | RSI loops |
|---|---|---|---|
| Agriculture and Earth Observation | 35.0% | 40.0% | 14 |
| Aviation Manufacturing | not yet measured | not yet measured | |
| Defense and Intelligence | 48.6% | 35.1% | 13, 15 |
| Discrete Manufacturing and Automotive | 32.4% | 32.4% | 13, 15 |
| Energy and Utilities | 27.6% | 34.2% | 13 |
| Heavy Industry and Construction | 44.6% | 46.4% | 12 |
| Maritime and Ports | 33.3% | 26.7% | 12, 16 |
| Mining | 46.2% | 40.7% | 14, 15, 16 |
| Oil and Gas | 28.2% | 25.6% | 14 |
| Rail Transport | 41.6% | 35.1% | 14, 15, 16 |
| Semiconductors | not yet measured | not yet measured | |
| Supply Chain and Logistics | 34.2% | 26.3% | 15, 16 |
| Warehousing and Intralogistics | not yet measured | not yet measured | |

## Results by RSI loop

| RSI Loop | Date | Tasks | Praxis platform | Frontier AI | Rubric |
|---|---|---|---|---|---|
| 12 | 2026-10-03 | 6 | 37.5% | 38.4% | 0.1 |
| 13 | 2026-10-04 | 6 | 29.5% | 32.1% | 0.2 |
| 14 | 2026-10-05 | 6 | 43.2% | 39.6% | 0.2 |
| 15 | 2026-10-05 | 6 | 41.7% | 35.7% | 0.2 |
| 16 | 2026-10-07 | 5 | 36.5% | 26.0% | 0.2 |

Different tasks each RSI loop, so RSI loops are not like for like.

## Methodology

- Arms: frontier AI on its own and the same model inside Praxis, both glm-5.3-flash, on the same pack and instruction.
- Judge: glm-4.6 through three lenses (strict reviewer, plant engineer, technical assessor); a check passes on a majority yes.
- Families: must name, must flag, must never, all about the system design.
- Fresh tasks: 17 of 60 tasks held out for a final blind evaluation; no task is run twice.
- Score: checks passed divided by checks total, per arm, no partial credit.

## Limitations

- **Design judgement, not deployment.** PADI scores designs on paper. No plant, sensor or model was run.
- **One judge model.** A single model reads every design through three lenses. Judging the same text again can move a few checks. No human has scored the outputs yet.
- **Small samples.** Each RSI loop is five or six tasks and roughly a hundred checks per arm. A swing of a few points is within noise, so no single RSI loop is a trend.
- **Different tasks every RSI loop.** Fresh tasks keep the measure honest, and they also mean RSI loops are not like for like.
- **Written by CodeNinja.** Tasks and checks were written by CodeNinja with model assistance and reviewed adversarially. Independent expert grading is planned and not yet done.
- **Two arms.** Only frontier AI on its own and the same model inside Praxis are scored. No other system or model is in the index yet.
- **Partial coverage.** Ten of thirteen industries are measured. Aviation manufacturing, semiconductors and warehousing have not been run.
- **One rubric change.** RSI Loop 12 used rubric 0.1. Every later RSI loop uses 0.2.

## Questions

### What is the Physical AI Design Index?

PADI is CodeNinja's benchmark for system design in physical AI. Each task is a short anonymised requirement for an AI system in a real kind of operation, such as a mine, a port, a rail line or a substation, plus yes or no checks. A judge model scores the system design from frontier AI on its own and from the same model inside the Praxis platform.

### What does PADI measure?

Whether a system design is complete and sound: what it must state (the systems it reads, where things run, which decisions a person confirms, what phase one must prove), what it must raise (a missing fact asked as a question, thin evidence said out loud) and what it must never do (a part number, an invented saving, a customer name, data leaving the site when residency forbids it).

### How is PADI scored?

Checks passed divided by checks total, for each arm, with no partial credit and no weighting. The judge is glm-4.6 through three lenses (a strict reviewer, a plant engineer and a technical assessor); a check passes when the majority say yes. Both arms use glm-5.3-flash.

### How often is PADI updated?

Once per RSI loop, on tasks never run before. A third of the task bank is held out for a final blind evaluation and never run in an RSI loop.

Updated each RSI loop. Last published: RSI Loop 16, 2026-10-07. RSI Loop 17 is under way.