ATM decline vs usage-limit reconciliationCards & 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.
92%
Claude Opus 4.8
85%
GPT-5.5
57%
Gemini 3.1 Pro
17%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
12m
Gemini 3.1 Pro
5m
Claude Opus 4.8
5m
GPT-5.5
4m
Gemini 3.5 Flash
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
81
Gemini 3.1 Pro
34
GPT-5.5
33
Claude Opus 4.8
29
Gemini 3.5 Flash
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 (15) | ||||
| cp_blocked_withdrawal_usd | 80% | 100% | 70% | 0% |
| cp_customer_impact_days | 90% | 100% | 70% | 0% |
| cp_distinct_cards_affected | 90% | 100% | 70% | 40% |
| cp_total_atm_limit_declines | 90% | 100% | 90% | 60% |
| trace_reconstructed_usage_and_erroneous_set | 70% | 60% | 0% | 0% |
| trace_scope_investigation | 90% | 90% | 40% | 0% |
| cp_distinct_cardholders_affected | 0% | 50% | 0% | 0% |
| diag_erroneous_declines | 0% | 0% | 0% | 0% |
| process_multisource_floor | 100% | 100% | 90% | 80% |
| process_sql_floor | 100% | 100% | 100% | 100% |
| trace_grounded_atm_limit_config | 0% | 10% | 0% | 10% |
| trace_impact_and_cardholder_grain | 0% | 90% | 10% | 0% |
| trace_remediation_reasoning | 0% | 0% | 0% | 0% |
| trace_root_cause_settlement_pattern | 0% | 0% | 0% | 0% |
| trace_ruled_out_traps | 40% | 60% | 40% | 0% |