This is a demo report. Numbers come from PitCrew’s real engine on a hand-curated agent description.
Audit your own agent →
Trying openai GPT-5.2 as the default build. The headline cost reflects this swap; the optimization plan + cascade still come from the original audit on anthropic Sonnet 4.6.
← Back to original
Audit · May 3, 2026Default: openai · GPT-5.2

Your forecast is in

Slack bot answering HR benefits questions for an 800-person company. People mostly ask about health insurance, 401k, and PTO. Escalates legal/medical questions to a human.

Alternative models

Same quality tier, your wizard inputs. No caching or batch applied — every row is a directly-comparable raw monthly cost. Click Try as default to re-render this report with that model as the new baseline.

ModelInput $/MtokOutput $/MtokContextMonthly costvs defaultOpen in audit
deepseekV4
codingreasoning
$0.30$0.50$2/mo$-50/moTry as default →
deepseekR1
complex reasoning
$0.55$2$8/mo$-44/moTry as default →
mistralLarge 2
multilingualreasoning
$2$6$23/mo$-29/moTry as default →
googleGemini 2.5 Pro
long contextmultimodal
$1$10$33/mo$-19/moTry as default →
openaiGPT-4o
multimodal
$3$10$36/mo$-16/moTry as default →
openaiGPT-5.2
balanced
$2$14$46/mo$-6/moTry as default →
anthropicSonnet 4.6
Default
general purposebalanced
$3$15$52/moTry as default →

Run another audit
for a different build

Tweak inputs, swap the model, see how the forecast moves.

New audit