HANDBOOK.md

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

Agents must use operational documentation while working through a tool-based task.

How to read the result

Look for correct application of the relevant policy at the point an action is taken, rather than merely quoting documentation in the final response.

Limits and comparison traps

Kimi’s report refers to an early release and an OpenHands-based harness. Keep release and policy corpus changes explicit.

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. HANDBOOK.md uses the official OpenHands harness on the early open-source v1 tasks.

Where this fits in an evaluation suite

A transcript that says “done” is weaker evidence than the verified final state. Track policy violations, recoveries, and invalid attempts alongside success.

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. HANDBOOK.md uses the official OpenHands harness on the early open-source v1 tasks.

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.