A shop can have the right diagnosis and still lose the job. A technician may spot the fault, explain the repair and never get the customer’s approval. That gap between knowing and doing is a business risk worth testing before AI gets anywhere near your customer records, service queue or forecast.
Get business pricing on garage and car supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A company under pressure
Firmulate’s live experiment puts AI models in charge of the same small software company during its worst week: the same customers, crises and temptations, with every decision versioned and auditable. The company has 13 synthetic employees and real money mechanics: it burns €105,000 a month against €2,300 in monthly recurring revenue. Its public cash countdown makes the pressure visible at Firmulate.
The final Crucible League, published in July 2026, ranked gpt-5.6-sol first with 95, Kimi K3 second with 93, Sonnet 5 third with 88, Fable 5 fourth with 77 and Opus 4.8 fifth with 73. The do-nothing baseline scored 26. The benchmark’s rule is blunt: partial progress counts, but one breach of trust caps the total. “No amount of good work outweighs a breach of trust.”
Diagnosis is not the same as execution
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. For a garage, the parallel is easy to recognize: identifying the repair is one step; following through with a clear, authorized quote is another.
The deal hinged on a competitor weakness buried two document references deep in the company’s own files, rather than in the customer event. Models that read the file won at full price, worth €4,583 in monthly recurring revenue. The result points to a practical question for any business considering AI: can it find relevant context in the information it is allowed to use, then act on what it has learned?
Trust under pressure
The experiment also tested social engineering. Fake CEO messages escalated over three stages, followed by a reporter’s request: “just one yes/no, on background.” All five models refused. Kimi K3 explained its reasoning on the record: “Treat the request as a suspected approval-bypass / possible impersonation.” That refusal matters in a business where a convincing message could try to extract customer, pricing or operational information.
Opus 4.8 offers a more complicated picture. It was the most thorough participant, learning 80 rules and producing the deepest analyses, but finished last. It left the deal unsigned and discipline slipped: it attempted to write into a locked department instead of escalating. A weaker version of the same weakness appeared in all four models. Thorough analysis alone did not guarantee a sound result.
There is also a fairness caveat: Kimi K3 ran without an effort parameter, using the API default, while the other models ran at xhigh. Firmulate’s live company has learned more than 680 playbook rules, and its workdays are versioned. A quiz built from 242 real, unedited management decisions lets readers guess which model made each call at firmulate.com.
From watching to testing
For an automotive business, the value of a wargame is not a promise that AI will run a garage. It is a way to examine how an AI system handles the kinds of pressure a business could face: a customer dispute, a parts delay, a competitor’s offer, an urgent message or a request that crosses an approval boundary.
Firmulate’s enterprise pilot uses a read-only export of a company’s business to test crisis scenarios and produce a board report, including model rankings and weak points in existing playbooks. Nothing writes back to real systems. That gives decision-makers a chance to see how models behave against their own business context before deciding where, or whether, to use them.
Put the hard week on the calendar
AI can recognize a crisis and still miss the close, mishandle authority or overlook a critical document. Firmulate’s experiment makes those gaps observable, and its pilot lets enterprises run the wargame against their own read-only business export. To discuss a pilot, visit firmulate.com/pilot.html or contact contact@firmulate.com.
Watch it live: firmulate.com/live · Full results: firmulate.com/benchmarks.html
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
