Choose by the revision you need
| Priority | A workflow to evaluate | The useful test |
|---|---|---|
| Different opinions on the same argument | Quartz’s document and model comparison | Can you explain which suggestion you kept and why? |
| Developing standalone content | Claude artifacts | Does iteration preserve the purpose of your draft? |
| Editing within a broad assistant | ChatGPT Canvas | Does the editing flow fit the other tasks you do? |
Scroll the table sideways to compare.
Sources:Anthropic: Claude artifactsOpenAI: ChatGPT capabilities
Ask for a diagnosis before a rewrite
‘Make this better’ invites changes without a shared standard. Say who will read it, what they should understand, and what cannot change. Ask where the draft falls short of that goal before requesting replacement text.
For an application, the problem may be an unclear example. For a recommendation, it may be an unsupported conclusion. A more elegant sentence will not repair either on its own.
A question to start with
Keep a record of your judgment
Read the original and the revision together. Mark any change in the claim, emphasis, or level of certainty. A good revision should solve a problem you recognize, not merely sound more professional.
In Quartz, you can select the useful part of a response and place it into your document. You can then edit the wording in context. Start with one passage and compare the result with the draft you had before.
Judge a finished paragraph
Try two tools with the same brief and passage. Give each one follow-up turn. Then compare the paragraph you would actually send, including the time you spent correcting it. You do not need a universal model ranking to discover a useful writing process.
A few practical questions
The short ones are here. For anything else, ask below.
Can I start using Quartz today?
Quartz is opening access through its waitlist. You can explore the public demo, then join the waitlist for your own workspace. Signing in is for people who already have access.
Does agreement between AIs mean the answer is correct?
No. Agreement can help you identify a common view, but models can share an error. Check important claims against original evidence. Different answers are a reason to investigate, not a vote you have to follow.
Which model writes best?
That depends on your task, brief, and standards. This page compares workflows; it does not claim a measured model-quality ranking.