Institutional fee-revenue reconstructionRevenue & Fees
How each model performed on this task. Every chart states its N, and the reward carries a 95% confidence interval once there is more than one trial to measure.
Core metrics
Check Pass %
Mean share of the task's checks passing, averaged over trials. N = 10 trials per model.
88%
Gemini 3.1 Pro
70%
GPT-5.5
70%
Claude Opus 4.8
14%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
4m
Gemini 3.5 Flash
3m
Gemini 3.1 Pro
3m
GPT-5.5
1m
Claude Opus 4.8
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
41
Gemini 3.5 Flash
21
Gemini 3.1 Pro
13
Claude Opus 4.8
10
GPT-5.5
Verifier × model pass rate
How often each verifier passed, per model. Green is passing, red is failing — a red band across a row is a verifier nothing clears.
| Verifier | GPT-5.5n=10 | Claude Opus 4.8n=10 | Gemini 3.1 Pron=10 | Gemini 3.5 Flashn=10 |
|---|---|---|---|---|
| checks (8) | ||||
| cp1_peak_year | 100% | 100% | 100% | 10% |
| cp2_peak_year_net | 0% | 10% | 40% | 10% |
| cp3_top_fee_category | 100% | 100% | 100% | 20% |
| cp4_top_category_net | 50% | 40% | 100% | 10% |
| cp5_oon_atm_launch_year | 100% | 100% | 100% | 20% |
| process_sql_floor | 100% | 100% | 100% | 40% |
| rubric_collected_not_scheduled | 0% | 10% | 40% | 10% |
| rubric_consulted_fee_policy | 40% | 100% | 40% | 100% |