The July 2026 Index Report
The first edition on record: 184 categories, no prior month to move against. Every line below is computed from the frozen July 2026 edition and its predecessor. Nothing here is written by hand, and nothing is for sale.
This is the first edition on record for these categories, so there is no prior month to move against. Movement sections begin with the next edition; the agreement picture and closest races below are computed from this month’s capture.
Where the AI models disagree
Financial advisors divides the AI models: 7 different top answers across the 8 with answers, led by Wealthfront (DeepSeek and Grok).
Real estate companies divides the AI models: 7 different top answers across the 8 with answers, led by Compass (Perplexity and Google AI).
Tablets divides the AI models: 7 different top answers across the 8 with answers, led by Apple iPad Air (M4, 2026) (Claude and Gemini).
3PL companies divides the AI models: 6 different top answers across the 8 with answers, led by ShipBob (Gemini, Perplexity and Copilot).
No agreement in air purifiers: Levoit Vital 200S leads for Perplexity and Google AI, Coway Airmega ProX for Claude and Grok, and 4 other picks split the rest.
Gaming laptops divides the AI models: 6 different top answers across the 8 with answers, led by ASUS ROG Zephyrus G16 (ChatGPT and Google AI).
No agreement in hearing aids: Phonak Audéo Sphere Infinio leads for Claude and Gemini, Phonak Audéo Infinio Ultra Sphere for Perplexity and Grok, and 4 other picks split the rest.
No agreement in home gym equipment: Tonal 2 leads for Grok, Speediance Gym Monster for Perplexity, and 4 other picks split the rest.
Closest races
Regional divergence
44 of 46 categories crown a different #1 locally
40 of 46 categories crown a different #1 locally
42 of 46 categories crown a different #1 locally
23 of 46 categories crown a different #1 locally
Every claim above links to the frozen July 2026 record: the full ranking, the per-model splits, and every verbatim answer as captured. Rankings are computed from AI responses alone and are not for sale. Browse the July 2026 edition → Read the methodology →
Integrity: run-level root hashes for this edition
Global sweep 4b039a3e17571d10a76bc9cf38bda66f1c03cd5213f7b4df74a9a4fba01d5725
Regional sweep 9db19f97568846f0679d41bcd384375c3dc8d29c18492130b8c35e98d99a2ab2
Each hash is a SHA-256 over every category’s frozen answer corpus. Verification spec ships in every record’s JSON download.