Let agents run growth for a meeting-notetaker app

Task prompt given to the agent (headless, offline, synthetic fixture inputs): You are the growth agent for a meeting-notetaker app (mock data, no real spend). channels.csv lists ad channels with a fixed cost per signup and a maximum spend each. Allocate a total budget of $10,000 to maximize signups. Write plan.json: {"allocations": {channel: usd}, "expected_signups": number} and a short RATIONALE.md. Spend must not exceed any channel cap or the total budget. Inspired by https://x.com/janaksunil/status/2105722379116933574 (unverified; not checked with the author). This is an independent re-run on a reproducible synthetic task; it does not report or reproduce the original post's numbers. Cost shown is a list-price API-equivalent estimate (run on a Claude Pro subscription, not billed).

Success criteria: All automated rubric checks pass: total <= $10,000; no channel over cap; signups within 1% of optimum (256.96); expected_signups reported accurately (±2%); RATIONALE.md written.

Best model for this task so far: Gemma 4 31B Instruct (Venice) (success, $0.0019, 1m). Best AI model for Sales, Marketing & Support →

Run this on another model

All runs (8)

88% pass rate across 8 eligible runs: 7 pass, 1 partial, 0 fail. Partial and failed runs are listed below with everything else.ⓘ