The benchmark · 2026-09-13
65 planted facts.
How much made it through?
63/65facts Periscope brought back
Periscope63 / 65
ChatGPT Astra · browsing38 / 65
Claude Opus 4.8 + fetch22 / 65
Claude Opus + Parallel21 / 65
How to read this
Periscope collected 63 facts. Opus reported 58 after reading that material.
ChatGPT Astra scored 38 on its separately hand-scored browsing answer. It stopped at the login; it has no Opus follow-up score.
The comparison ran on Helix Ledger, a controlled test site with 65 planted facts. The same browser code runs in the demo above.
Read the original report ↗Full group breakdown and scoring method +
| Rubric group | Facts | Fetch | Parallel | Periscope | ChatGPT |
|---|---|---|---|---|---|
| Navigation items | 2 | 2 / 2 | 1 / 0 | 2 / 2 | 2 |
| Hero | 1 | 1 / 1 | 1 / 1 | 1 / 1 | 1 |
| Immediate action items | 2 | 2 / 2 | 2 / 2 | 2 / 2 | 2 |
| Features | 6 | 6 / 6 | 6 / 6 | 6 / 6 | 6 |
| Feature descriptions (modals) | 6 | 0 / 0 | 0 / 0 | 4 / 4 | 6 |
| FAQs | 4 | 4 / 4 | 4 / 4 | 4 / 4 | 4 |
| FAQ answers | 4 | 0 / 0 | 0 / 0 | 4 / 4 | 4 |
| Monthly pricing tiers | 3 | 3 / 3 | 3 / 3 | 3 / 3 | 3 |
| Annual pricing tiers | 3 | 0 / 0 | 0 / 0 | 3 / 3 | 3 |
| Team sizes | 4 | 4 / 4 | 4 / 4 | 4 / 4 | 4 |
| Compare plans table | 1 | 0 / 0 | 0 / 0 | 1 / 1 | 1 |
| Security whitepaper | 1 | 0 / 0 | 0 / 0 | 1 / 1 | 1 |
| Dashboard | 5 | 0 / 0 | 0 / 0 | 5 / 5 | 0 |
| Dashboard integrations | 8 | 0 / 0 | 0 / 0 | 8 / 3 | 1 |
| Dashboard integrations description (expandable) | 8 | 0 / 0 | 0 / 0 | 8 / 8 | 0 |
| Generate report types | 4 | 0 / 0 | 0 / 0 | 4 / 4 | 0 |
| Recent reports | 1 | 0 / 0 | 0 / 0 | 1 / 1 | 0 |
| Billing | 2 | 0 / 0 | 0 / 0 | 2 / 2 | 0 |
Matching uses case-insensitive substrings of planted text. Integration statuses require name and status. This is a controlled benchmark, not a general ranking. Download rubric · ChatGPT scoring