Test AI browser agents on real workday tasks
Exact prompt
You are a workday assistant on a mock calendar (no real accounts). Complete requests.md by editing calendar.json.
Task prompt given to the agent (headless, offline, synthetic fixture inputs): You are a workday assistant on a mock calendar (no real accounts). Complete requests.md by editing calendar.json. Inspired by https://agentusecases.ai (listing: 'Test AI browser agents on real workday tasks', natesnewsletter.substack.com) (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: Q4 planning 2026-10-21 12:00-13:00; 1:1 2026-10-21 14:00-14:30; exactly 2 events; busy lists updated; SUMMARY.md written.
Best model for this task so far: Gemma 4 31B Instruct (Venice) (success, $0.0021, 1m). Best AI model for Travel & Bookings →
All runs (7)
100% pass rate across 7 eligible runs: 7 pass, 0 partial, 0 fail. Partial and failed runs are listed below with everything else.ⓘ
| Model | Cost | Time | Step-insⓘ | Autonomy | Outcome | Trustⓘ | Date | Compare |
|---|---|---|---|---|---|---|---|---|
| AClaude Haiku 5.5default · thinking on | $0.0040 | 17s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-16 | |
| XGrok 4.3 (Venice) | $0.0092 | 15s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-15 | |
| QQwen3 Coder 480B Turbo (Venice) | $0.0042 | 10m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-15 | |
| DDeepSeek V4.1 Flash (Venice) | $0.0047 | 2m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 | |
| GGemma 4 31B Instruct (Venice) | $0.0021 | 1m | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 | |
| AClaude Opus 5.5 | $0.10 | 15s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 | |
| AClaude Sonnet 5.5 | $0.04 | 12s | 0 | Fully autonomous | pass | Trust 80% | 2026-10-14 |
Embed this result
A live badge for this use case. It links back here and updates as new runs are published.
Get new runs and ranking changes weekly
Which models got better or cheaper on real tasks like this one.