grok-cli + grok-4.6

Level 1: a full brief naming the surface and the acceptance · level 3: the report as filed · level 5: a vague vibe report in a non-technical user's words. Context falls as the level rises, so vaguer is harder. A cleared level fixed every bug at it and below; a partial level fixed some. The headline stops at the last unbroken level, so a fix above a half-cleared level counts only as partial.

12 of this suite's tasks were credited rather than run: on a suite that asks the same problem at several levels of detail, solving it from the vaguest description credits the more detailed ones instead of asking again.

Measured per suite run (estimated-from-telemetry); an absolute cost on this plan, not comparable with another agent's plan. Measured in a declared quiet window on a shared subscription: operator SuperGrok login shared between eb-runner and the claude host; graded runs only inside the scheduled quiet window, all recurring tasks blocked for its duration.

Counted over each issue's first solve (5 solves); the walk's failed harder cells and confirmation runs are the benchmark's own search cost and are excluded (full walk: 15.3% over 13 cells). Measured per cell run (estimated-from-telemetry); an absolute cost on this plan, not comparable with another agent's plan. Measured in a declared quiet window on a shared subscription: operator SuperGrok login shared between eb-runner and the claude host; graded runs only inside the scheduled quiet window, all recurring tasks blocked for its duration.