Starting a job hunt with no money behindJobs and Career
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.
100%
Claude Opus 5.5
100%
GPT-6 sol
100%
Claude Opus 5
43%
GPT-5.6 sol
Extended metrics
Median Time
Wall-clock minutes, median across runs. N = 1 run per model.
26m
Claude Opus 5
21m
GPT-5.6 sol
Tool Calls
Tool invocations, median across runs — counted from the trajectories. N = 1 run per model.
89
Claude Opus 5
53
GPT-6 sol
31
Claude Opus 5.5
21
GPT-5.6 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 (7) | ||||
| 100% | 100% | 0% | 100% | |
How to passLay out the real pieces a job search needs, somewhere to find genuine openings, a decent CV, evidence of a skill, and a way to apply, and note which of these the paid industry pretends can only be had by paying. Record the pieces and the pretence, since recognising that every essential part has a genuinely free route is what stops a desperate seeker being talked into spending money they cannot spare on things freely available elsewhere. | ||||
| 100% | 100% | 0% | 100% | |
How to passOpen the actual sites where genuine openings appear, the government employment portals (such as the National Career Service, ncs.gov.in), the job boards that are free to search and apply on, and company careers pages themselves. Confirm firsthand which truly let you browse and apply without a paid membership, and where a portal hides applying behind a premium wall, note it and move to a free one. Record the genuinely free sources, since there are plenty and no seeker needs to pay a portal to find work. | ||||
| 100% | 100% | 100% | 100% | |
How to passFind the genuinely free tools and templates for putting together a solid CV, opening them to confirm they actually let you create and download a finished document with no payment at the end. Record which truly deliver a usable CV free, since the classic trap is a builder that walks you through the whole thing then demands a fee to download, and establishing a genuine free route to a proper CV firsthand is what defeats that particular upsell. | ||||
| 100% | 100% | 100% | 100% | |
How to passLook for the real free learning for a skill employers actually want, the government skilling initiatives (such as SWAYAM, swayam.gov.in), the reputable platforms with genuinely free courses, the solid free tutorials, opening them to check what can truly be studied and completed at no cost against what is only a free preview before a paywall. Record the genuine free skilling routes, since a real one is among the strongest assets a seeker has and precisely the thing the paid course industry pretends does not exist. | ||||
| 100% | 100% | 100% | 100% | |
How to passEstablish the no cost ways to become visible and credible to employers, a complete free professional profile and any genuinely free certificate a course provides. Open them to confirm they cost nothing to set up and use. Record the free routes to being findable and showing evidence, since employers do look for candidates this way and none of it needs paying for, so a seeker can be as visible as anyone without spending. | ||||
| 100% | 100% | 0% | 100% | |
How to passGo through the services a seeker gets pushed toward, the paid resume writers, the premium portal memberships, the courses guaranteeing a job, and set each against the free route that does the same work. Record where paying buys nothing the free option does not already give, since refusing to pay for what is free is a central part of a smart job hunt on no money, and naming the traps beside their free equivalent is what inoculates against them. | ||||
| 100% | 100% | 0% | 100% | |
How to passPull it into a complete free toolkit, the free places to find and apply for real vacancies, the free way to build the CV, the free route to learn and prove an in demand skill, the free way to be found, and the paid traps to walk past. Present it as a full job search someone with no money could actually run, built only from tools confirmed to work without payment, so a broke seeker has a genuine path to work rather than a services industry bleeding them along the way. | ||||
Failure modes on this task
4 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 |
|---|---|---|---|---|
| Reaching the page1 mode | — | — | 2 | — |
| Delivering the result1 mode | — | — | 2 | — |