Ask four AIs the same question.
On local spots, they don’t agree.
What is the best burger in SF?
OpenAI
Zuni Cafe
Claude
In-N-Out Burger
Gemini
Causwells
Perplexity
Hamburger Project
Across 20 picks, 15 different names. Nothing makes every list.
See the full rankingGallop lines up the top fives from OpenAI, Claude, Gemini and Perplexity and measures the overlap. Cloud providers, they nearly match. Burgers in San Francisco, not even close.
Ask your own questionEvery question, from least to most agreement
What is the best burger in SF?
#1 picks: Zuni Cafe, In-N-Out Burger, Causwells, Hamburger Project
Little overlapTop coffee shops for remote work?
#1 picks: Blue Bottle Coffee, Local independent cafes, Stumptown Coffee Roasters
Little overlapBest programming language for startups?
#1 picks: Python, JavaScript
Mostly agreeWhich AI company will lead in 2025?
#1 picks: Anthropic, Nvidia
Mostly agreeBest design tools for 2025?
#1 picks: All four say Figma
Mostly agreeMost reliable cloud provider?
#1 picks: All four say AWS
Same five
How it works
Each model gives a ranked top five for the same question. Gallop lines the lists up and compares them, nothing more.
Agreement is the average overlap between each pair of top fives. Same five in any order counts as full agreement. Names that only differ in spelling, like Blue Bottle and Blue Bottle Coffee, are merged first.
The rankings come from Gallop’s first prototype and were saved in one batch, labeled by month. How they were made wasn’t recorded: not the prompts, not the model versions, not whether each month was asked in that month. Read it as a sample, not a study. Nobody checked whether any answer is right.
Lists labeled Jul 2024 to Jul 2025.