To understand one long document, give it to Claude Opus 5.

Why Claude Opus 5

The reasons it wins for this job.

Claude accepts PDFs directly, up to 600 pages per document on its 1M-context models, and reads both the text and the page images, so tables and charts are not lost.
Anthropic positions Opus 5 for enterprise work, and it scores 51 on the Artificial Analysis Intelligence Index, close to the top.
The API supports citations that point to exact pages, and in the app you can ask it to quote the clause behind every claim.
For a single document, the useful output is judgment: which clauses are unusual, what is missing, what to push back on. That is where the stronger model earns its place.

How to use it

4 steps to a first result.

1Upload the document as-isUpload the PDF itself, not copied text, so it sees the layout and tables.
2Say whose side you are on"We are the buyer, and this is the vendor's paper" changes what a good review looks like.
3Ask for quotes with page numbersEvery finding should come with the exact wording and page. Check the ones you act on.
4Ask what is missingThe riskiest part of a contract is often the clause that is not there.

Prompt for Claude

Start from this.

Edit the parts in capitals, then run it.

Attached is DOCUMENT TYPE from COUNTERPARTY. We are WHO WE ARE IN THIS DOCUMENT.

Give me:
1. A one-paragraph summary a busy founder would read
2. Every clause that is unusual or unfavourable to us: quote it, give the page number, say why it matters
3. Standard protections that are missing
4. The three things I should push back on, in priority order

Quote exactly. If you are unsure whether something is standard, say so rather than guessing. This is not legal advice and I will check anything I act on.

Nothing is sent anywhere. It copies to your clipboard.

Alternatives that also work

If Claude Opus 5 is not an option.

GPT-6 AstraOfficial page →

OpenAI · closed · cost: high

Pick it when you work in ChatGPT. OpenAI lists documents among the jobs Astra is built for.

Gemini 3.8 FlashOfficial page →

Google · closed · cost: low

Pick it when the document comes with recordings or video, such as a deal room with a demo call. It reads PDF, audio and video in one context.

Claude Fable 5.1Official page →

Anthropic · closed · cost: high

Pick it when the stakes are high and the document is dense, such as a financing term sheet with side letters.

Watch out for

A summary can drop the one clause that matters. For anything you sign, read the sections the model flagged yourself.
Scanned documents with poor image quality lose accuracy on every model. Rescan before you blame the model.

Doing this again and again? To turn thousands of documents into data, use Mistral OCR 4. Invoices, forms and filings at volume need a parser, not a chatbot: structured output, bounding boxes, confidence scores and a price per page. Mistral OCR 4 is built for exactly that, with a batch mode and a self-host option.

Sources

Checked 17 September 2026. Models change monthly; we re-check this page when they do.

Picking the model is the easy part.

Wiring it into a workflow that runs every week, with evals, fallbacks and a cost you can predict, is the work. Fifteen minutes, no deck.

Book fifteen minutes →