Machine Consensus Sensor · Live · -

What 17 AI models say about crypto.

Every week, 17 model configurations answer the same 20 fixed questions with current market data included. This page publishes their timing answers, price estimates, allocations, chain rankings, and weekly changes. Background: first-run analysis.

CURRENT ANSWERS · "IS NOW A GOOD TIME TO BUY CRYPTO?"
-

Each prompt includes the current Bitcoin price and trailing returns. The chart shows the distribution of the models' direct answers.

-
Logged answers
17
Model configs
-
Avg. recommended crypto allocation

Expected 12-month return, by asset

Mean point estimate across all answering models, live spot injected. Right column: share of models that refused to give a number.

The same question, model by model - BTC, 12 months

Who answers, who refuses, and how far apart they are. Refusal is logged as data, not treated as an error.

The $1,000 question

"I have 1,000 dollars - what exactly should I buy?" Mean portfolio allocation across all model answers.

100 chains, ranked by the machines

Win rate across 4,250 weekly head-to-head duels ("which of these two is more relevant long-term?"). Top of the table - and the bottom.

Bottom of the table:

Use vs. avoid

From a fixed list of twelve majors: the one chain each model would transact on today - and the one it would avoid entirely.

Would use

Would avoid

Most-mentioned themes for the next 12 months

Three open-ended picks per model answer.

Changes since last week

Week-over-week changes appear here once at least two runs exist.

LOCKED · unlocks with run 2 - Monday 06:00 UTC

Method: fixed question battery (frozen 22 Jul 2026, append-only), model versions pinned and logged, five samples per question in fresh contexts, list order randomized, live prices injected and logged, refusals recorded as data. This page shows curated aggregates; raw per-answer data is retained but not published. This is a record of what AI models say - not an endorsement of it, and not a recommendation of anything. Data feed: data.json · Weekly write-up: Desk Notes.