A BRXND project · built by Alephic

GEO Leaderboard

Who is winning the AI visibility game? A point-in-time leaderboard of the GEO platforms most often recommended by leading LLMs — measured by what the models themselves say.

Run
2026-07-06
LLM responses
1,225
Prompt panel
175
01 - This week's headlines
Leader

Profound leads with 42% mention rate, 31% first-named.

2.5× the SoV of #2 Peec AI. Dominant on ChatGPT (GPT-5.5) (54%) and Gemini 3.1 Pro (49%); strongest in category leader and jargon-aware prompts.

Punching above weight

AthenaHQ ranks #5 on only $2M raised — outranking Brandlight ($30M raised, 13.6× the war chest) and 2 other heavier-funded peers.

Funding tier: the smallest war chest among the 7 consistently-visible companies. Mention rate 11% — outpunching peers with materially more capital behind them.

Opportunity gap

Implicit (no tool ask) prompts are open territory.

Even the strongest player — AthenaHQ at 2% — hasn't really cracked them (vs Profound at 70% in category leader). Whoever wins this slice unlocks share no incumbent currently holds.

02 - Share-of-Voice leaderboard

Click any company to see where they win

Company
SoVcomposite
Mention rate% of responses
03 - How to read this
SoV score
A composite relative score (not a percentage) combining mention rate + first-mention bonus + ordinal-position decay, weighted by audience-weighted surface multipliers and prompt category. Higher = more prominent in the LLM answers a real CMO actually sees.
Mention rate
% of LLM responses where the company appears at all. Each response can mention several companies — so mention rates do not sum to 100%.
First %
% of responses where the company is the first name in the answer. A strong signal that the LLM treats them as the category default.
Avg mention order
When the company is mentioned, the ordinal index of their first appearance in the response text (1.0 = first named). Not a search ranking — LLM answers are prose with embedded lists. (n=K) flags fewer than 10 backing mentions.

See the full methodology for the scoring formula, surface audience weighting, statistical significance bounds, and what we explicitly don’t claim.

04 - Surfaces tracked
Claude Sonnet 4.6Gemini 3.1 ProGoogle AI OverviewsChatGPT (GPT-5.5)Perplexity Sonar-Pro

All 5 run with live web search enabled

05 - Explore

Snapshot from run 2026-07-06 · methodology v2.0