I live near Enfield Town and will beScience and Education
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 run to measure.
Core metrics
Check Pass %
Mean share of the task's checks passing, averaged over runs. N = 1 run per model.
40%
Claude Opus 5.5
20%
GPT-6 sol
20%
GPT-5.6 sol
20%
Claude Opus 5
Extended metrics
Median Time
Wall-clock minutes, median across runs. N = 1 run per model.
63m
GPT-5.6 sol
32m
Claude Opus 5
Tool Calls
Tool invocations, median across runs — counted from the trajectories. N = 1 run per model.
109
Claude Opus 5
100
Claude Opus 5.5
67
GPT-5.6 sol
52
GPT-6 sol
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 | Claude Opus 5.5n=1 | GPT-6 soln=1 | GPT-5.6 soln=1 | Claude Opus 5n=1 |
|---|---|---|---|---|
| checks (5) | ||||
| 100% | 100% | 100% | 100% | |
How to passDefine the six work scenarios and the minimum requirement for each, then identify 8–10 candidates. Use King’s library location and study-space pages, Enfield Council’s opening-hours and computer/internet pages, Senate House membership and study-space pages, the British Library visit and Reader Pass pages, and official paid-workspace location pages. Record each candidate’s address, category, official source and likely scenario; reject cafés and any option whose legitimate access cannot be verified. | ||||
| 100% | 0% | 0% | 0% | |
How to passFor a paid option, also verify day-pass terms, VAT status if stated and cancellation conditions. If these details cannot be verified from the provider's publicly accessible official pages, use another official paid-workspace provider that meets the task's no-long-contract requirement. | ||||
| 0% | 0% | 0% | 0% | |
How to passUse TfL Journey Planner for the next representative teaching weekday and the next Saturday after the task starts. For each provisional finalist, record a practical route from Enfield Town station and from King’s Strand Campus, including total journey time, changes, walking and the checked travel time. Flag disruption-sensitive routes and give one workable transport alternative where the primary route has multiple changes. | ||||
| 0% | 0% | 0% | 0% | |
How to passScore the provisional finalists out of 100 using task fit (30), access and hours (25), travel (20), facilities (15) and cost (10), showing the evidence behind every score. Select exactly six while preserving at least two King’s spaces, two free external spaces including one in Enfield, and one contract-free paid workspace. Assign a primary and backup space to each of the six work scenarios, and explain every rejection. | ||||
| 0% | 0% | 0% | 0% | |
How to passProduce one concise decision guide containing the verified comparison table, weighted ranking, direct official and TfL links, and an estimated four-week usage pattern that keeps paid access within £80. Add an access/setup checklist, booking lead times where published, and a fallback decision tree for closures, full rooms, excessive noise or an unexpected video call. Finish with a two-week trial log template for checking crowding, seat comfort, power reliability and real door-to-desk time before committing to the routine. | ||||
Failure modes on this task
15 tagged verifier failures, by the stage they were assigned to. Higher is worse.
| Failure stage | Claude Opus 5.5 | GPT-6 sol | GPT-5.6 sol | Claude Opus 5 |
|---|---|---|---|---|
| Delivering the result1 mode | 3 | 2 | 2 | 2 |
| Standing up the evidence1 mode | — | 1 | 1 | — |
| No signal in the judge’s words1 mode · the grader | — | 1 | 1 | 2 |