Batch a regulatory briefing into one call

01 · Role in the system

What Jev does here

Jev answers eight Noul, two Choice, and three Score questions against the same 53,777-character document. The cookbook repeats both batching strategies five times and compares each answer's tracked probability or normalized score. Local code pins and fetches the article, caches calls, calculates variance, tokens, cost, and elapsed time, and renders the comparison.

The economics of sharing one long document

Thirteen typed questions — eight Noul, two Choice, three Score — against a pinned 53,777-character GDPR article, answered in a single call and compared against thirteen sequential calls across five runs. The batched strategy came out 12.2 times cheaper, and the reason is structural: when a long shared document dominates every request, sending it once is the whole win.

The cookbook is careful about its own headline. The 10.0x latency advantage sums sequential single-call latencies; concurrent requests would shrink that gap without changing the token story. And workloads with little shared state get little of this benefit, which the limitations state directly.

What makes this a usable reference

Everything in the comparison is measured by code: variance across runs, per-answer probabilities, tokens, cost, elapsed time. The document is pinned and cached so the comparison runs against the exact same text, and both strategies use the same model version, jev-1.12.

For anyone structuring document-heavy judgment workloads, this is the reference record: batch independent questions over shared context, keep scores comparable across strategies, and let the numbers — not the architecture diagram — argue for the design.

02 · Control boundary

Where Jev sits

A document-dominated workload batches independent typed judgments so shared state is sent once; code measures and compares the two call strategies.

Code owns the loop, permissions, thresholds, validation, and side effects. Jev owns only the bounded judgments described above.

03 · Known limits

What this evidence does not prove

• The cost advantage is largest because the long shared document dominates each request; workloads with little shared state will differ.
• The reported speed ratio sums sequential single-call latencies, and concurrent requests would reduce that latency gap without reducing repeated input tokens.

04 · Attribution

Public sources

This is an Official record: the example was published by TypeSafe itself.

• TypeSafe ↗Official · cookbook · public · checked 2026-09-18