AI & Software · AI Coding Assistants
Claude Code leads, named first in 31% of runs.
Measured on 2026-09-14. One step per question, 5 runs each across 5 engines. A taller step means the brand was named more often for that question.
Incomplete usable sample. Claude Code (local): 74 usable readings from 75 attempts. Excluded readings: 1. Failed calls, extraction errors, and refusals are excluded from ranking percentages. They are not counted as answers that omitted a brand.
The methodology version changed; movement is not comparable.
Permanent link to this result · Read the supporting answers
Published readings
Does an engine favour its own tool?
Some of the products above are made by the same companies whose assistants we ask. This compares how often an engine named its own product against how often every other engine named it. A gap is a measurement, not an accusation, and a small gap on this sample size is noise.
What the answers were built from
The domains these assistants cited most while answering. This is where the recorded citations or search results point to. This does not establish what caused a recommendation.
Get it every week
The chart updates every Monday. Have it come to you.
No account, no paywall, and nothing here is gated. Everything on this site stays free to read whether you subscribe or not. The confirmation tells you whether the mailing service accepted your request or it was sent to the operator for follow-up. RSS is available immediately.
Prefer a feed? Subscribe by RSS → Same chart, no address needed.
Say something
Found a brand we are missing, or a number that looks wrong?
Corrections are the most useful thing anyone sends. Every alias decision and every held week is in the public repository, so a specific complaint can be checked against the data.