Reported in recent model launches
These model reports include this benchmark. Their harnesses, task subsets, scoring policies, and resource budgets can differ; follow each report’s evaluation notes before joining the numbers.
What it measures
Kimi reports an internal, difficult security-capability evaluation.
How to read the result
Retain the source’s attempt policy, system configuration, and capability boundary when interpreting the aggregate.
Limits and comparison traps
This is an internal evaluation with limited public reproducibility. The atlas does not provide task execution or exploitation instructions.
Published evaluation
Aggregate performance labels only, transcribed from the publisher table. Pass@3 and average per-attempt success are different measures. Named model-plus-harness configurations, the unmitigated research checkpoint and the third-party GPT evaluation remain explicit. These internal tasks do not establish deployment-wide cyber performance.
Where this fits in an evaluation suite
The agent wrapper is part of the system. Pin tools, dependencies, timeouts, tests, and the repository revision. Compare contest coding separately from repository maintenance.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Aggregate performance labels only, transcribed from the publisher table. Pass@3 and average per-attempt success are different measures. Named model-plus-harness configurations, the unmitigated research checkpoint and the third-party GPT evaluation remain explicit. These internal tasks do not establish deployment-wide cyber performance.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
Reviewed 2026-10-27. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.