Home / For vendors
Private testing for the teams who build AI app builders.
We run your builder, and the competitors you choose, through the same app briefs and rules as our public benchmark, privately, with recordings and repro steps. What we never sell is a public score.
What you can get
Free
Right of reply
Before a result changes, we send your team the scorecard, the repro steps and the screenshots. Tell us within 7 days if something is factually wrong and we'll fix it publicly.
$0
One-off
Private competitive benchmark
Your builder plus up to three competitors, on our standard app briefs (SlotKeeper today; more as they're published), three runs each, with ChatGPT, Claude, Grok and Gemini on the same task as a baseline, on free and cheapest paid plans, with screen recordings, timings, a bug list with repro steps, and a written report. Delivered in 10 business days.
$1,500
Founding rate: $750 for our first three clients.
Monthly
Release retest
We rerun the briefs after each major release and send you a private trend line, so regressions show up before your users find them.
$500 /month
The rules (they apply to you too)
- Nobody can pay for a score or a placement. Not now, not later.
- Every tool gets the same prompt, the same rules and the same rubric.
- We publish what we didn't test, next to what we did.
- Mistakes go in a public corrections log, including our own inconsistencies.
- If a tool we test ever buys private testing from us, its public page will say so.
Buying private testing doesn't change your public score, its timing, or its wording.
What a private report contains
- The same receipts as our public pages, for every run.
- Screen recordings and timings for each run.
- Every bug with severity and repro steps.
- Rubric scores per run, plus the spread across three runs.
- Human cleanup minutes for every run: how much work your users still do after your tool says it's finished.
- How your tool does against ChatGPT, Claude, Grok and Gemini on the same task: the "do you even need the specialist tool?" question, answered for your product.
- A plain-language summary of where your builder wins and loses against the competitors you picked.
How to start
Send a message to @BenchCoded on X with your tool's name. We'll reply with a written scope. No calls needed.
Message @BenchCoded on XSee the evidence we send
NotePrices are our launch prices and may change. Nobody has bought private testing yet; when someone does, their public page will say so.