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.

Kimi · Kimi K3.1 Pro →DeepSeek · DeepSeek V4.1 Flash 1017 →Meta · Muse Wonder and Muse Wonder Full →

What it measures

Agents reconstruct programs from observed behavior rather than receiving a conventional repository repair issue.

How to read the result

Distinguish a graded reconstruction score from a binary threshold applied to that score. Reproducing most behavior and passing an all-or-nothing gate answer different questions.

Limits and comparison traps

Launch reports sometimes threshold the score at 95 percent. That binary result must not be pooled with the original graded metric.

Published evaluation

The linked publication documents the task and scoring procedure. Model results retain the source’s evaluation setting, native score units and available uncertainty.

Published evaluation

Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation.

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

Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation.

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

  1. Introducing Kimi K3.1 Pro ↗

    Moonshot AI · 2026

  2. ProgramBench — evaluation methodology and results ↗

    Vals AI · 2026

Reviewed 2026-10-27. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.