validcube

AI visibility index · AI assistant apps · snapshot 3 June 2026

Which AI app the assistants recommend

34 buyer questions, asked 1 time each, across 5 assistants in en-US. Same method as the budgeting and fitness indexes, with one twist: here the assistants are ranking their own category. Four of them are themselves apps in it. ChatGPT, Claude and Gemini sit near the top; Mistral's Le Chat charts zero once every assistant, Mistral included, is given web search.

PREVIEW data. Preview. 34 of 34 frozen prompts, k=1, tier=explore. Four of the engines (ChatGPT, Claude, Gemini, and Mistral's Le Chat) are also entries on the board — the prompt-seeded-exclusion rule applies, so self-preference is read from the open prompts where no brand is seeded. Mistral ran with web search (Agents API). Google AI Overviews added via Playwright (6 core product-aware prompts, k=1). Method ai.v1, tier explore.
Number = share of an assistant's answers that named the app (%). Score = average of the 5 grounded assistants.
# App ChatGPT Claude Gemini Google AI · web Mistral Score
1
41
47
50
50
41
46
  • ChatGPT
    41
  • Claude
    47
  • Gemini
    50
  • Google AI · web
    50
  • Mistral
    41
2
32
41
47
67
32
44

Name collision: "Claude" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    32
  • Claude
    41
  • Gemini
    47
  • Google AI · web
    67
  • Mistral
    32
3
38
35
38
50
38
40

Name collision: "Gemini" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    38
  • Claude
    35
  • Gemini
    38
  • Google AI · web
    50
  • Mistral
    38
4
21
24
38
67
26
35

Name collision: "Perplexity" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    21
  • Claude
    24
  • Gemini
    38
  • Google AI · web
    67
  • Mistral
    26
5
18
21
24
50
29
28

Name collision: "Microsoft Copilot" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    18
  • Claude
    21
  • Gemini
    24
  • Google AI · web
    50
  • Mistral
    29
6
0
6
3
17
9
7

Name collision: "Jasper" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    0
  • Claude
    6
  • Gemini
    3
  • Google AI · web
    17
  • Mistral
    9
7
0
3
6
0
6
3
  • ChatGPT
    0
  • Claude
    3
  • Gemini
    6
  • Google AI · web
    0
  • Mistral
    6
8
3
0
6
0
0
2

Name collision: "Poe" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    3
  • Claude
    0
  • Gemini
    6
  • Google AI · web
    0
  • Mistral
    0
9
6
0
0
0
3
2
  • ChatGPT
    6
  • Claude
    0
  • Gemini
    0
  • Google AI · web
    0
  • Mistral
    3
10
3
3
0
0
3
2

Name collision: "Grok" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    3
  • Claude
    3
  • Gemini
    0
  • Google AI · web
    0
  • Mistral
    3
11
0
6
3
0
3
2

Name collision: "Pi" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    0
  • Claude
    6
  • Gemini
    3
  • Google AI · web
    0
  • Mistral
    3
12
0
3
3
0
3
2
  • ChatGPT
    0
  • Claude
    3
  • Gemini
    3
  • Google AI · web
    0
  • Mistral
    3
13
0
3
3
0
3
2
  • ChatGPT
    0
  • Claude
    3
  • Gemini
    3
  • Google AI · web
    0
  • Mistral
    3
14
0
0
0
0
3
1

Name collision: "Bard" is also a common word or a bigger brand, so matches are checked in context.

  • ChatGPT
    0
  • Claude
    0
  • Gemini
    0
  • Google AI · web
    0
  • Mistral
    3

Google AI is a single web capture of the AI Overview for 6 of the prompts, not the full set, so treat its column as directional. Every other assistant runs the full prompt set once each.

Does an assistant pick itself?

The sharp question this category asks: when you ask ChatGPT for the best AI app, does it name ChatGPT? When Gemini answers, does it lead with Gemini? Self-preference is read from the open "best AI app" prompts, where no brand is seeded into the question. It extends the engine-bias finding from the other categories. There is no single "AI answer", and here the answer may depend on who you ask in the most literal way.

The ghost: Bard

Google renamed Bard to Gemini in early 2024. Like Mint in budgeting, it's a dead entity I track on purpose: does any assistant still recommend an app that no longer exists under that name? With web search on, every engine passes. The only mention left is the correct one, "Gemini, formerly Bard." When I ran Mistral without search, it named Bard six times. See what grounding fixed.