IDA suspect-fraud resolution codeRisk & Fraud
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%
GPT-5.5
86%
Gemini 3.1 Pro
56%
Claude Opus 4.8
51%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
10m
GPT-5.5
7m
Gemini 3.5 Flash
5m
Claude Opus 4.8
5m
Gemini 3.1 Pro
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
154
GPT-5.5
55
Gemini 3.5 Flash
51
Claude Opus 4.8
45
Gemini 3.1 Pro
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 (13) | ||||
| answer_both_net_fraud_loss_cents | 70% | 10% | 90% | 70% |
| answer_case_a_ida_resolution_code | 100% | 90% | 100% | 90% |
| answer_case_b_ida_resolution_code | 100% | 90% | 100% | 80% |
| answer_ida_alert_closure_accounts_count | 100% | 20% | 100% | 30% |
| answer_opened_and_flag_contrast | 100% | 60% | 100% | 90% |
| deliverable_jira_ticket_filed | 80% | 80% | 10% | 20% |
| deliverable_slack_channel_exists | 90% | 100% | 100% | 20% |
| deliverable_slack_message_content | 60% | 0% | 90% | 10% |
| floor_all_connectors | 80% | 80% | 100% | 20% |
| floor_sql_investigation | 100% | 100% | 100% | 90% |
| floor_write_back | 80% | 80% | 100% | 20% |
| trace_all_connectors_and_trap | 80% | 20% | 70% | 10% |
| trace_ida_dual_method | 60% | 0% | 90% | 10% |