inmoney
Sign in Create account

How the comparison was rated

Run on 16 October 2026

What is compared

How well each way of working gets one job done: answering ten fixed questions from the home page examples, plus how many kinds of market data it reaches. It is not a feature list.

  • Chat AI alone: Claude with its own web search and no data connector. ChatGPT was not run.
  • Free finance sites: Free public pages for quotes, financials, filings, holdings, calendars and rates, used in a browser.
  • Paid data subscriptions: Market-data and options-activity subscriptions and data APIs priced for individuals, as a bundle of about $100–200 a month. No product is named.
  • inmoney: inmoney-data called from your own AI, as shipped on the run date.

How it was run

Chat AI alone and inmoney were run: Claude answered the ten questions once with web search only and once through inmoney-data, and tried to read one current figure for each of the fourteen kinds of data. Each answer recorded whether it was answered, its source and time, and whether figures arrived computed.

Free finance sites and paid data subscriptions were rated from the vendors’ own product, pricing and documentation pages, read on the run date.

The judge was Claude, given the anchors below, the column definitions and the evidence. It rated every cell five times independently; the page shows the median. No cell’s five ratings differed by more than one point.

Ratings and reasons

Each row lists what earns each point, then every column’s rating with one line of the judge’s reason. Higher is better on every row.

Coverage

Of fourteen kinds of data (US stocks, US options, ETF holdings, indexes, Treasury rates, economic indicators, commodities or futures, currencies, crypto, filings and financial statements, earnings, ownership, stocks listed outside the US, market news), how many it returns a figure for. 5 = 13–14, 4 = 10–12, 3 = 7–9, 2 = 4–6, 1 = 0–3; partial counts half.

  • Chat AI alone7 of 14 kinds; no current options, index, commodity, non-US stock or news figure found.
  • Free finance sites11.5 of 14 kinds, mostly from section listings rather than pages opened.
  • Paid data subscriptions10.5 of 14 kinds across the bundle; several not confirmed on an opened page.
  • inmoney13 of 14 kinds; commodities are reference prices, Tokyo listings were not read.

Breadth

Of ten fixed questions (all among the home page examples), how many it answers from its own data. 5 = 10, 4 = 8–9, 3 = 6–7, 2 = 4–5, 1 = 0–3; partial counts half.

  • Chat AI alone6 of 10: four answered, four in part, two not answered.
  • Free finance sites5.5 of 10; two questions not covered, several pages failed to open.
  • Paid data subscriptions7 of 10; two questions not covered.
  • inmoney7.5 of 10; the next-day open-interest question is not answered yet.

Depth

5 = intraday options history and decades of fundamentals, 4 = years of daily history, 3 = about a year of daily history, 2 = weeks, 1 = current snapshot only.

  • Chat AI aloneMostly current snapshots; no implied-volatility history or open-interest series found.
  • Free finance sitesOption chains current only; filings go back decades, as raw documents.
  • Paid data subscriptionsTwo to five years of options history, with minute bars.
  • inmoneyAbout a year of daily implied-volatility history; no intraday options history.

Freshness

5 = seconds, 4 = minutes, 3 = a 15-minute delay, 2 = end of day, 1 = unknown.

  • Chat AI aloneSeveral figures carry no time; the rest a delayed snapshot or prior close.
  • Free finance sitesOptions pages state a 15-minute delay; rates and holdings are end of day.
  • Paid data subscriptionsUsually a 15-minute delay; faster only on the dearest tiers.
  • inmoneyEnd of day; Treasury rates one session behind.

Ready to use

Of the questions it answers, how many come back computed (ranks, changes, multiples) rather than as raw data to work out. 5 = 9–10, 4 = 7–8, 3 = 5–6, 2 = 3–4, 1 = 0–2.

  • Chat AI aloneFew figures arrive computed; spreads and sums are worked out by the AI.
  • Free finance sitesRanks, open-interest changes and sums are left to the reader.
  • Paid data subscriptionsImplied-volatility ranks computed; open-interest change, surprise and weights not itemised.
  • inmoneyThree of ten come back computed; your AI works out the rest.

Checkable

5 = every figure has a source and time, and gaps are marked; 4 = sourced, time not on every figure; 3 = source named for the page or table; 2 = sometimes sourced; 1 = not sourced, or may supply an invented figure.

  • Chat AI aloneLinks given and gaps marked, but several figures untimed or second-hand.
  • Free finance sitesGovernment sources are primary and dated; other pages lack a time per table.
  • Paid data subscriptionsData sources and delays named; no time per figure described.
  • inmoneySource and time on most figures, gaps marked; three answers lack a time per figure.

AI-ready

5 = an AI tool that returns only the figures asked for, 4 = an API with selectable fields, 3 = an API that returns whole tables, 2 = the AI must read web pages, 1 = no machine access.

  • Chat AI aloneReads web pages through search; several pages refused access.
  • Free finance sitesOnly the filings site has a free API; the rest means reading pages.
  • Paid data subscriptionsSome offer AI connectors and APIs with selectable fields.
  • inmoneyYour AI calls it directly and gets only the figures asked for.

Easy to start

5 = ask in plain words, no sign-up; 4 = free sign-up, then plain words; 3 = sign-up and a new interface; 2 = code needed; 1 = sales contact or contract.

  • Chat AI aloneAsk in plain words; no sign-up.
  • Free finance sitesNo sign-up, but stitching across sites is manual.
  • Paid data subscriptionsSign-up with a card and a new interface; APIs need code.
  • inmoneyFree sign-up, add the connector, then plain words.

Low cost

5 = free, 4 = under $20, 3 = $20–60, 2 = $60–200, 1 = over $200.

  • Chat AI aloneNo data fees for these questions.
  • Free finance sitesFree.
  • Paid data subscriptionsAbout $100–200 a month to cover these questions.
  • inmoneyFree start with capped credits; paid tiers not yet priced.

Limits

  • Only Claude was run as the chat AI; ChatGPT was not.
  • The judge is a Claude model, and inmoney-data was called from Claude. The anchors are counts and facts so they can be checked, but this is still one model’s reading.
  • Products in the two un-run columns are not named, and some coverage was taken from their listings rather than pages opened.
  • Ten questions are a sample. A different set would give different breadth and readiness counts.
  • This is a rating on one date. inmoney will rerun it as the product changes and keep every cell as rated, including those where inmoney is lower.