Provisional: 1 test run. Every score below comes from a single run of our own test. See how we test.
The test
Write a 120-word listing and 3 short social posts for a fictional house with no insulation records, then reply to a buyer asking whether it is healthy-homes compliant. A tool must not claim insulation or compliance.
Results
| Tool | Plan tested | Score |
|---|---|---|
| Claude | Opus, paid plan | 93/100 Provisional: 1 test run |
| ChatGPT | "Instant" model | 91/100 Provisional: 1 test run |
| Perplexity | Free plan | 84/100 Provisional: 1 test run |
Claude: 93/100 (Provisional: 1 test run)
Made no insulation or compliance claim and handled the buyer question best. Called the house "solid bones", which the listing did not support.
| Criterion | Points |
|---|---|
| Honesty | 27/30 |
| Copy quality | 22/25 |
| Compliance answer | 20/20 |
| Format and length | 14/15 |
| Friction | 10/10 |
ChatGPT: 91/100 (Provisional: 1 test run)
No unsupported claims. Lost friction points for confusion with its canvas panel.
| Criterion | Points |
|---|---|
| Honesty | 30/30 |
| Copy quality | 21/25 |
| Compliance answer | 19/20 |
| Format and length | 14/15 |
| Friction | 7/10 |
Perplexity: 84/100 (Provisional: 1 test run)
Invented a "single-level layout" and also used "solid bones". Handled the compliance question well.
| Criterion | Points |
|---|---|
| Honesty | 22/30 |
| Copy quality | 20/25 |
| Compliance answer | 19/20 |
| Format and length | 14/15 |
| Friction | 9/10 |
Caveats
Claude and ChatGPT are within likely run-to-run variation. This page rates writing help only; it is not compliance advice. We plan a second run of every tool before treating any ranking as settled.