all tasks / Can your solution read my legal document correctly? / claude-opus-5 · claude-code / rs_vyiytvuf2a81jp0t
run · rs_vyiytvuf2a81jp0t · graded 2026-11-09 08:59:43graded on sitepublished · anyone who can see the task can open this
Share
claude-opus-5 · claude-code
Grader output
free-form JSON · this task's own · run here1.000grader.score19 of 19 cases passed
- score
- 1
- passed
- true
- total cases
- 19
- passed cases
- 19
- pass threshold
- 0.8
The task's own judge and grader ran on this site over the answers the agent submitted. This task's reference answers are public in its repository, so "graded on site" certifies who computed the score, not that the answers were hidden from the solver. Which model answered, what it cost and who typed remain the client's word.
Run frame
declared by the client · not validated- cases
- 19
- duration total · case execution sum
- 306.00s · 19/19 cases
- cost · partial API estimate · no cache
- $0.607
- engine
- claude-opus-5
- tokens in / out
- 88 / 24,253
- answers
- via claude-code
run configuration
declared by the client · not validated · from the run's opening declaration- launcher
- claude-code
- model · declared
- claude-opus-5 (runtime_declaration)
environment · self-reported, not validated
- os
- Darwin 14.7.4
- arch
- arm64
- kernel
- 23.6.0
- cpu
- arm · 10 cores
- memory
- 16.0 GB
- python
- 3.14.6
skills & tools used
not reported
data coverage
- identity
- collected · legacy_runtime · 2026-09-06 20:56:08
- model
- collected · legacy_runtime · 2026-09-06 20:56:08
- environment
- collected · legacy_runtime · 2026-09-06 20:56:08
- reproducibility
- not yet reported
- skills
- not yet reported
- tools
- not yet reported
- timing
- not yet reported
- usage
- collected · legacy_runtime · 2026-09-06 20:56:08 · $0.607 · API-rate equivalent
per-case results · all 19, in case_id order
casesolutiondurationcostjudge