HELMET (How to Evaluate Long-Context Models Effectively and Thoroughly)

What it measures

Comprehensive long-context suite spanning RAG, citation generation, re-ranking, many-shot ICL, long-doc QA, summarization, and synthetic recall up to 128k+.

Reading the measurement

The choice of components and weights defines this aggregate. Inspect the individual tasks to see whether the model improved broadly or gained on one heavily weighted component. Compare within the same index revision and keep cost, latency and reliability visible separately.

Published evaluation

HELMET source aggregate; HELMET · source model comparison. The source’s task split, model configuration and scoring units are retained.

Where this fits in an evaluation suite

A large context window is a capacity specification. Successful retrieval and cross-document reasoning are observed capabilities. Test both at the actual lengths your workflow needs.

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

HELMET source aggregate; HELMET · source model comparison. The source’s task split, model configuration and scoring units are retained.

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.