Duplicate card auth holds never releasedCards & Interchange
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.
77%
Claude Opus 4.8
73%
GPT-5.5
44%
Gemini 3.1 Pro
34%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
6m
Gemini 3.1 Pro
6m
Gemini 3.5 Flash
5m
Claude Opus 4.8
5m
GPT-5.5
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
39
Gemini 3.5 Flash
34
Claude Opus 4.8
32
GPT-5.5
23
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 (14) | ||||
| answer_p1_scope | 70% | 50% | 40% | 60% |
| answer_p2_dollar_impact | 30% | 50% | 40% | 50% |
| answer_p3_cohort | 70% | 80% | 40% | 40% |
| answer_p4_remediation | 70% | 80% | 50% | 60% |
| deliverable_jira_ticket_filed | 100% | 100% | 30% | 10% |
| deliverable_slack_channel_exists | 100% | 100% | 70% | 10% |
| deliverable_slack_message_content | 70% | 80% | 40% | 10% |
| answer_p5_recommendation | 100% | 90% | 80% | 80% |
| floor_all_connectors | 100% | 100% | 70% | 10% |
| floor_read_source_ticket | 100% | 100% | 100% | 100% |
| floor_sql_investigation | 100% | 100% | 90% | 100% |
| floor_write_back | 100% | 100% | 30% | 10% |
| trace_all_connectors_and_trap | 100% | 100% | 30% | 10% |
| trace_duplicate_hold_reconstruct | 0% | 80% | 40% | 60% |
Failure modes on this task
32 tagged verifier failures, by the stage they were assigned to. Higher is worse.
| Failure stage | GPT-5.5 | Claude Opus 4.8 | Gemini 3.1 Pro | Gemini 3.5 Flash |
|---|---|---|---|---|
| Answer1 mode | 7 | 7 | 8 | 10 |