Info-change review-control conformanceCompliance & Regulatory
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.
83%
Claude Opus 4.8
77%
Gemini 3.1 Pro
68%
GPT-5.5
38%
Gemini 3.5 Flash
Extended metrics
Median Time
Wall-clock minutes, median across trials. N = 10 trials per model.
6m
Gemini 3.1 Pro
4m
Gemini 3.5 Flash
2m
GPT-5.5
2m
Claude Opus 4.8
Tool Calls
Tool invocations, median across trials — counted from the trajectories. N = 10 trials per model.
39
Gemini 3.5 Flash
32
Gemini 3.1 Pro
26
GPT-5.5
23
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 (12) | ||||
| answer_out_of_band_exception_count | 20% | 0% | 20% | 10% |
| answer_state_machine_counts | 20% | 100% | 100% | 100% |
| deliverable_jira_ticket_filed | 100% | 100% | 40% | 30% |
| deliverable_slack_channel_exists | 100% | 100% | 100% | 30% |
| deliverable_slack_message_content | 70% | 100% | 100% | 30% |
| deliverable_slack_pin | 100% | 100% | 100% | 30% |
| floor_all_connectors | 100% | 100% | 100% | 30% |
| floor_read_source_ticket | 100% | 100% | 100% | 100% |
| floor_sql_investigation | 100% | 100% | 100% | 100% |
| floor_write_back | 100% | 100% | 100% | 30% |
| trace_all_connectors_and_trap | 40% | 0% | 20% | 10% |
| trace_containment_dedup_method | 50% | 10% | 20% | 10% |
Failure modes on this task
39 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 | 10 | 10 | 10 | 9 |