Repository changes face tests and a maintainer-focused rubric, with blocking failures receiving no credit.
28 model results
One row per model and compatible score measure. Mean = equal-weight arithmetic average of imported reports. Distinct configurations reported for the same model in a source table remain separate. Reports can overlap; the range is not a confidence interval. Unbundle to inspect each observation.
Publisher-selected recent-model cohort from the public evaluation page. Model labels preserve effort and fallback configurations. Values are a dated snapshot, not a live feed. Consult the publisher’s methodology for the harness, task split and judging; do not merge this cohort with vendor-run results.
TRACE THE SCORE · Per-model notes use published breakdowns where available.
What it measures
Can an agent deliver a change that survives repository tests and maintainer review? Partial rubric credit distinguishes incomplete work from a fully acceptable patch.
Version boundary
This is the 2.0 task set. FrontierCode 1.1 results remain on their own page; a percentage from one release is not a score on the other.
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
Artificial Analysis: 132 private tasks, three attempts each, Stirrup, 500 turns, GPT-6 Luna (medium) judge. Blocking-check failures and solution-lookup flags score zero; task scores receive equal weight.
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.