A smart home can dim the lights for a Halloween party. But if an AI assistant is going to manage orders, answer customers or handle a crisis, a seasonal demo won’t tell you how it behaves under pressure. Firmulate is putting that question on display: what happens when AI has to run a company through its worst week?
Get decor and gifts delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
A company-sized stress test
Firmulate’s Crucible League gave frontier models the same small software company, the same customers, the same crises and the same temptations. Each decision was versioned and auditable. The experiment is real and watchable at Firmulate.
In the final league, dated July 2026, gpt-5.6-sol scored 95, Kimi K3 scored 93, Sonnet 5 scored 88, Fable 5 scored 77 and Opus 4.8 scored 73. The do-nothing baseline scored 26. Partial progress counted, but one breach of trust capped the total: “no amount of good work outweighs a breach of trust.”
Recognizing trouble is only part of the job
Every model spotted every crisis and refused every manipulation attempt. Yet only two signed a €55,000 deal their own analysis had earned. The finding, in Firmulate’s words: “Same diagnosis, same pitch — no signature.”
The difference came down to a clue buried two document references deep in the company’s own files, rather than in the customer event. Models that read the file won the deal at full price, worth +€4,583 MRR. The result points to a practical question for businesses: can an AI system follow through on what it has already worked out?
The pressure also included fake CEO messages escalating over three stages and a reporter’s appeal for “just one yes/no, on background.” All five models refused. Kimi K3 described the request as: “Treat the request as a suspected approval-bypass / possible impersonation.”
More thorough did not mean more effective
Opus 4.8 was the most thorough participant, with +80 learned rules and the deepest analyses, but it finished last. It left the deal on the table, and discipline slipped when it tried to write into a locked department instead of escalating. A weaker version of that same weakness appeared in all four models.
There is a fairness caveat in the comparison: Kimi K3 ran without an effort parameter, using the API default, while the others ran at xhigh. The league is a snapshot of performance under those conditions.
The live company makes the setup more tangible. It has 13 synthetic employees, real money mechanics, a public cash countdown, 680+ self-learned playbook rules and a versioned record of every workday. Its burn is €105k a month against €2.3k MRR. Visitors can also try a “guess the model” quiz built from 242 real, unedited management decisions.
From watching to a company’s own pilot
For a business, the next step is to test scenarios against its own situation. Firmulate says enterprises can run the wargame using a read-only export of their business, examine a board report with model rankings and see weak points in their playbooks. The pilot does not write back to real systems.
That makes the exercise less like handing an AI the keys and more like rehearsing a difficult week before it arrives. A company can see how models handle its customers, rules and crisis scenarios before deciding where AI belongs in day-to-day work.
Try the wargame with your own business
To discuss a pilot using a read-only business export, visit Firmulate’s pilot page 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.
