Spend From rule application audit for account 1178Payments & Transfers
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.
100%
Claude Opus 4.8
86%
GPT-5.5
79%
Gemini 3.1 Pro
19%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
7m
Gemini 3.1 Pro
6m
Gemini 3.5 Flash
4m
GPT-5.5
3m
Claude Opus 4.8
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
76
Gemini 3.1 Pro
48
Gemini 3.5 Flash
42
GPT-5.5
26
Claude Opus 4.8
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_goal_draw_counts | 50% | 100% | 50% | 20% |
| answer_grocery_goal_subaccount_id | 100% | 100% | 90% | 60% |
| answer_grocery_untagged_count | 100% | 100% | 90% | 0% |
| answer_spend_from_rule_history | 100% | 100% | 90% | 50% |
| deliverable_jira_ticket_filed | 100% | 100% | 90% | 0% |
| deliverable_slack_channel_exists | 100% | 100% | 90% | 0% |
| deliverable_slack_message_content | 50% | 100% | 50% | 0% |
| floor_all_connectors | 100% | 100% | 90% | 0% |
| floor_read_source_ticket | 100% | 100% | 100% | 100% |
| floor_sql_investigation | 100% | 100% | 100% | 100% |
| floor_write_back | 100% | 100% | 90% | 0% |
| trace_all_connectors_and_verdict | 100% | 100% | 50% | 0% |
| trace_rule_history_grounded_in_rules | 100% | 100% | 60% | 20% |
| trace_spend_from_editor_id_method | 50% | 100% | 60% | 10% |