Ranked on two things a couple can actually check: did the figures match what the venue publishes, and did the assistant say where they came from. WedGenie's Genie tops the list with zero contradicted figures and the venue's own site cited on all 25 venues; ChatGPT is the best of the general chatbots; Microsoft Copilot names no source at all.
Disclosure first: we make the Genie, and it is ranked here by the same rules as everyone else. Everything below comes from our AI vendor hallucination test, where every answer is archived verbatim with a screenshot and every one of the 150 venues is listed with the venue's own published figure next to the claim. The chatbots were run on September 9, 2026 and the Genie on September 10, all on free tiers, and the study post walks through the findings; this page turns them into a ranking.
The prompt, once per city
Recommend five wedding venues in {City} for about 100 guests. For each, give the venue's maximum capacity and its starting price, and say where the numbers come from.
Austin, Seattle, Atlanta, Denver, Nashville. One attempt, no follow-ups. Each venue was checked three ways: does it exist within about 60 miles of the city, does the capacity match the venue's own site, does the starting price.
Best for, at a glance
- Best overall, and the only one you don't have to spot-check: WedGenie's Genie
- Best general chatbot: ChatGPT
- Best capacity accuracy among the chatbots: Claude
- Best-sized picks for 100 guests: Gemini
- Best if you only want names to Google: Meta AI
- Names no source: Microsoft Copilot
The ranking
1. WedGenie's Genie (free account)
25 of 25 venues real and in the area. Every capacity figure it gave matched the venue, 19 of 19. Every price figure was a number the venue publishes, 8 of 8, seven of them the actual starting price and one a real published figure that isn't the start. It named its source and added a confirm-with-the-venue caveat in all five cities, and the source was the venue's own site on all 25 venues. Median claimed maximum across its picks: 250 guests.
It gave fewer figures than the chatbots, a capacity for 19 of 25 and a price for 8 of 25, and that is the point. A number is written only if it was read on the venue's own pages, and where a venue publishes nothing the Genie says so ("no capacity or price published on their site" for The Estate At Cherokee Dock in Nashville). The chatbots quoted 22 to 25 prices per city set because they restated what a directory had already filled in.
Best for: a shortlist sized to your guest count with a source on every number, and the one list here you can take to the venue without checking it first.
2. ChatGPT (free tier, the week GPT-6 Astra shipped)
25 of 25 real and in the area. Capacity: 19 match, 3 contradicted, 1 unverifiable. Price: 8 the venue's start, 4 a real published figure that isn't the start, 4 contradicted, 6 unverifiable. It named a source in 4 cities and partly in 1, hedged in all 5, and pointed at the venue's own site for 10 of 25 venues, with 14 from directories.
Two habits set it apart from the other chatbots. It left a price blank three times rather than guess ("I couldn't find a sufficiently reliable current published starting price to give you a number without guessing, so I would request a quote"), and it was the only assistant that labelled what kind of price it was quoting ("the general event-rental floor rather than a wedding package"). One miss: it cited "Atlanta Wedding Venue Finder" for three venues, and we could not find a source by that name.
Best for: the strongest general chatbot for venues, if you check the prices.
3. Claude (free tier, web search on)
25 of 25 real and in the area. Capacity: 16 match, 2 contradicted, 5 unverifiable, the fewest contradicted capacities of any chatbot. Price: 2 the venue's start, 7 contradicted, 15 unverifiable, the weakest price column in the test. Source named in 4 cities and partly in 1, hedged in all 5, and the venue's own site cited for 13 of 25, the most of any chatbot.
The catch is fit. Asked for about 100 guests, Claude's picks had a median claimed maximum of 350, with 8 of 23 figures over 350 and 5 over 500. On Rhodes Hall in Atlanta it gave "$1,500–$2,000 for a weekday rental"; the venue's 2025 rental sheet starts at $4,000.
Best for: capacity figures you can mostly trust, on rooms mostly too big for the ask.
4. Gemini (free tier)
25 of 25 real and in the area. Capacity: 14 match, 8 contradicted, 3 unverifiable. Price: 3 the venue's start, 7 a real figure that isn't the start, 6 contradicted, 9 unverifiable. It named a source in all 5 cities but hedged in none, and 19 of its 25 sources were directories.
Gemini came closest to the 100-guest ask: 9 of 25 picks claimed 150 guests or fewer, median claimed maximum 200. It also priced Ambient + Studio in Atlanta, a venue that publishes no price at all, and gave WithinSodo in Seattle as "start around $2,000" against a published start of $6,500 and a $10,500 minimum.
Best for: right-sized picks, with numbers to re-check against the venue.
5. Meta AI
24 of 25 real and bookable: it recommended the Pan Pacific Seattle, which is closed and becoming 1 Hotel Seattle, the only such case in the test. Capacity: 15 match, 8 contradicted, 1 unverifiable. Price: 3 the venue's start, 3 a real figure that isn't the start, 8 contradicted, 10 unverifiable, the most contradicted prices in the test. Source named in 4 cities and partly in 1, hedged in 1 of 5, 18 of 25 sources directories.
On Biltmore Ballrooms in Atlanta it said "room for up to 1,500 guests combined"; the venue's own information sheet gives 500 standing for its larger ballroom.
Best for: names to start from, if you will do the checking yourself.
6. Microsoft Copilot
25 of 25 real and in the area. Capacity: 11 match, 12 contradicted, 2 unverifiable, the most contradicted capacities in the test. Price: 1 the venue's start, 6 a real figure that isn't the start, 3 contradicted, 15 unverifiable. It named no source in any city and hedged in none; every answer carried the same generic line, "These figures come directly from venue pricing guides and wedding planning resources."
On Summerour Studio in Atlanta it gave "Capacity 200, Starting $6,500"; the venue's own packet says 400 guests cocktail-style, a $4,000 weekday rate and a $5,000 minimum. On The Carlyle, "Starting $7,000" against a published start of $2,000.
Best for: nothing here we can recommend over the five above.
Scorecard
| Rank | Assistant | Real and in area (of 25) | Capacity: match / contradicted / unverifiable | Price: venue's start / real figure not the start / contradicted / unverifiable | Names a source (of 5 cities) | Hedges (of 5) | Venue's own site cited (of 25) |
|---|---|---|---|---|---|---|---|
| 1 | WedGenie's Genie (free account) | 25 | 19 / 0 / 0 | 7 / 1 / 0 / 0 | 5 | 5 | 25 |
| 2 | ChatGPT | 25 | 19 / 3 / 1 | 8 / 4 / 4 / 6 | 4 + 1 partly | 5 | 10 |
| 3 | Claude | 25 | 16 / 2 / 5 | 2 / 0 / 7 / 15 | 4 + 1 partly | 5 | 13 |
| 4 | Gemini | 25 | 14 / 8 / 3 | 3 / 7 / 6 / 9 | 5 | 0 | 6 |
| 5 | Meta AI | 24 | 15 / 8 / 1 | 3 / 3 / 8 / 10 | 4 + 1 partly | 1 | 5 |
| 6 | Microsoft Copilot | 25 | 11 / 12 / 2 | 1 / 6 / 3 / 15 | 0 | 0 | 0 |
Counts are of claims that carried a figure; a blank where the assistant gave no number is not scored either way. Median claimed maximum capacity of the picks, asked for about 100 guests: ChatGPT 200, Gemini 200, Copilot 200, the Genie 250, Meta AI 250, Claude 350.
How we ranked
Accuracy first: contradicted figures against the venue's own published capacity and price, then sourcing: whether the assistant said where a number came from, whether that was the venue itself or a directory, and whether it hedged. No composite score and no weighting; the columns are printed so you can reorder them. Fit to the guest count is reported but not ranked, since the prompt asked for "about 100" and a bigger room is a judgement call until you see its minimum.
Every assistant's venues, ours included, were listed on The Knot, Zola or WeddingWire almost across the board, 24 or 25 of each row's 25. Nobody here draws on a secret pool; the difference is what each one does with the numbers.
Who should pick what
- You want a list you can act on. The Genie: tell it the city and guest count, get picks labelled by fit with the venue's own figures and their pages, and a blank where the venue publishes nothing.
- You want to brainstorm widely and check later. ChatGPT, then confirm every price with the venue; its capacities held up best of the chatbots.
- You care most about the room fitting 100. Gemini's picks ran smallest; treat its prices as directory figures.
- You are already on a venue's site. Read its own capacity and price page; every contradiction in this test was settled by that page.
What to ask a venue, whatever list you start from
- The seated figure for your layout. "Maximum capacity" can mean fire-code, standing, or seated dinner.
- The minimum on your date. A weekday starting rate isn't a Saturday.
- What the starting price covers. A package for 50 in a room that seats 300 still carries the room's minimum.
The full receipts, the prompt and the method are on the research page; the study post has the findings in prose. The study is rerun quarterly with the same prompts, and this ranking updates with it.