Software, tested on the same tasks as its rivals.
Each tool gets the same fixed prompts, input files and written pass rules. We save every answer, a screenshot and the date. You can check each result, or run the test yourself.
Each tool gets the same fixed prompts, input files and written pass rules. We save every answer, a screenshot and the date. You can check each result, or run the test yourself.
| Tool | Version tested | Passed | Score | Run date |
|---|---|---|---|---|
| Notion AI | Free plan, signed in | 9/10 | 9.0/10 | 2026-11-21 |
| ChatGPT | Free, not signed in | 8/10 | 8.0/10 | 2026-11-21 |
| Google Gemini | Free, not signed in | 8/10 | 8.0/10 | 2026-11-21 |
| Claude | Free plan | Not run | Not yet tested | Not run |
| Grammarly | Free plan | Not run | Not yet tested | Not run |
| Microsoft Copilot | Free plan | Not run | Not yet tested | Not run |
| Perplexity Comet | Free plan | Not run | Not yet tested | Not run |
| Raycast AI | Free plan | Not run | Not yet tested | Not run |
Score: tasks passed out of 10, from saved results only.Open the full results and every saved answer
Each procedure has a version. When a rule changes, older results stay visible but stop counting in scores.