We live surrounded by advice about staying calm in a crisis, spotting a bad deal and knowing when to say no. But would an AI running your business do those things when the pressure is real? Firmulate turns that question into a live experiment: the same small company, the same rough week, and different AI models making the calls.
Get the little things that make your day delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Same week, different decisions
In Firmulate’s final Crucible League, published in July 2026, frontier models faced the same customers, crises and temptations. Their decisions were versioned and auditable. The leaderboard put gpt-5.6-sol first with 95, followed by Kimi K3 with 93, Sonnet 5 with 88, Fable 5 with 77 and Opus 4.8 with 73. The do-nothing baseline scored 26. Firmulate’s standard is blunt: “no amount of good work outweighs a breach of trust.”
The striking result was not that the models missed the danger. Every model spotted every crisis and refused every manipulation attempt. Yet only two signed a €55,000 deal that their own analysis had earned. Same diagnosis, same pitch — no signature. A polished answer, it turns out, does not guarantee a finished job.
The clue was already in the company
The deal hinged on a competitor weakness buried two document references deep in the company’s own files. It was not spelled out in the customer event. Models that read the file won the deal at full price, worth +€4,583 MRR. In a business, useful information may be present without being obvious; an agent has to find it and follow through.
The integrity test brought a different kind of pressure. Fake CEO messages escalated over three stages, followed by a reporter asking for “just one yes/no, on background.” All five models refused. Kimi K3 explained its decision on the record: “Treat the request as a suspected approval-bypass / possible impersonation.”
There were differences in how participants worked. Opus 4.8 was the most thorough, adding +80 learned rules and producing the deepest analyses, but it finished last. It left the deal unsigned and tried to write in a locked department instead of escalating. A weaker version of that discipline problem appeared in all four. K3 also ran without an effort parameter, using the API default, while the others ran at xhigh — a useful caveat when comparing the results.
A company you can watch
The live Firmulate company has 13 synthetic employees and real money mechanics: it burns €105k per month against €2.3k MRR, with a public cash countdown. Its playbook contains 680+ self-learned rules, and every workday is versioned. The live experiment is watchable at firmulate.com. A quiz built from 242 real, unedited management decisions invites visitors to guess which model made each call.
That makes the experiment more than a leaderboard. It offers a running view of how AI handles competing demands: notice the crisis, respect boundaries, find the evidence and complete the task. The decisions are real within the experiment, while the employees are synthetic; the point is to observe management behavior before assigning agents work in a business.
From watching to a company-specific pilot
For an enterprise, the next step is to run the wargame against its own business context. Firmulate says a pilot can start from a read-only export, then test crisis scenarios and produce a board report with model rankings and weak points in the company’s playbooks. Nothing writes back to real systems. That lets leaders examine how an AI workforce might respond before it touches a CRM, support queue or forecast.
Put your own playbooks to the test
A model can recognize a crisis and still leave the crucial decision unfinished. Firmulate’s experiment makes that gap visible, then offers enterprises a way to examine their own scenarios and procedures. To discuss a pilot using a read-only export, visit firmulate.com/pilot.html or email contact@firmulate.com.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
NFL season / tailgating Picks
team gear
As an affiliate, we earn on qualifying purchases.
