Prompt
Pull facts from an image: the prompt, and its price
A prompt that reads a photo or screenshot you attach and puts its facts in a table. Copy it or open it in llmwise, and see which models can run it and what one run counts as on each.
Models and prices checked . Updated .
Short answer
Claude Haiku 5.5 is the least expensive model that can run it: one run counts as 1 everyday message. 19 of the 21 models can run it; each one's price is below.
The prompt
From the attached file, make a table of [the items: dates, amounts, names], with the page each one comes from. If something is unclear, say so instead of guessing.
Replace the parts in [brackets] with your own. It's in the prompt library with the others.
What it needs from a model
An image you attach. It counts as 1,600 tokens toward the chat's length, and only models that read images can use it: DeepSeek V4 Pro and GLM 5.3 can't.
Which models, and what one run counts as
Start with the least expensive model that can run it, and move up only when the answer isn't good enough: the next model sees the whole chat.
| Models | One run on a plan | Runs on Plus | Runs on Pro |
|---|---|---|---|
| Claude Haiku 5.5, GPT-6 Luna, GLM 5.3 Flash, and DeepSeek V4.1 FlashReads images · Everyday model: from the daily count | 1 everyday message | 30 a day | 60 a day |
| Mistral Large 4, Gemini 3.8 Flash, Claude Haiku 4.5, GPT-6.1 Terra, and Grok 4.7Reads images | 1 message of the model | 125 a month | 250 a month |
| Claude Sonnet 5.5, Claude Sonnet 5, GPT-6.1 Sol, GPT-6 Sol, Gemini 3.1 Pro (preview), and Kimi K3Reads images | 1 message of the model | 62 a month | 125 a month |
| Claude Opus 5.5 and Gemini 4 ArgonReads images | 1 message of the model | 31 a month | 62 a month |
| Claude Fable 5.1 and GPT-6 AstraReads images | 1 message of the model | Not on Plus | 31 a month |
| DeepSeek V4 Pro and GLM 5.3 | Can't run it: doesn't read images | ||
Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.
Everyday models: for formula help, quick lookups and explaining a chart.
Models with 125 to 250 messages a month on Pro: for cleaning data, pivots, standard statistics and writing the analysis code.
Models with 31 to 62 messages a month on Pro: for messy data, ambiguous questions and analysis you'll present or decide on.
What to check in the reply
Before you use the reply, read it against the source you gave it; these are the places this prompt's answers go wrong.
Compare the table against the picture cell by cell for the figures that matter; a smudged 3 and 8, or a 1 and 7, read alike to a model.
Check that it kept units and currencies that were in the photo, and didn't invent a column the picture never had.
Where the photo cuts off or blurs, expect gaps: ask it to mark what it couldn't read rather than fill it in.
How many images a chat takes
Up to 10 files a message, images up to 5 MB each. Every image stays in the chat and counts toward its length, so a chat full of screenshots comes to count as more per reply.
| Images in the chat, up to | A reply counts as |
|---|---|
| 40 | 1 message |
| 80 | 2 messages |
| 125 | 4 messages |
Questions
Can I run this prompt for free?
No: there's no free plan or trial. It runs on every plan. Paid monthly from the day you start. Cancel in two clicks in Settings → Billing. On Pro ($20 a month) the least a run counts as is 1 everyday message, on Claude Haiku 5.5.
Which models can run it?
19 of the 21 models in llmwise can. DeepSeek V4 Pro and GLM 5.3 can't: doesn't read images. Start on the least expensive and move up if the answer isn't good enough; the next model sees the whole chat.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.
See what a message costs before you send it. Every plan is month to month: cancel in two clicks. Sign in with an email link, no password.