AssistantBench

What it measures

214 realistic time-consuming web tasks (real estate, travel, shopping, local services) requiring synthesis across multiple pages.

Reading the measurement

The useful outcome is a verified change or deliverable. Inspect the tool interface, initial state, user simulator and policy constraints. An agent can claim completion without reaching the required state, or reach it through an action the protocol forbids.

Published evaluation

Accuracy; AssistantBench · main evaluation. Model configurations and missing observations follow the original publication.

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

Accuracy; AssistantBench · main evaluation. Model configurations and missing observations follow the original publication.

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.