JevMade

Sign in
← Back to Jev News Benchmarks & Research

Head-to-Head Benchmarks Show Where OpenAI Decisions, Jev, and Parallel Split

A November specification and benchmark comparison maps pricing, request limits, and leaderboard gaps between TypeSafe Jev, OpenAI’s Decisions API, and Parallel.

2 min read By JevMade Source: Parallel ↗
Editorial collage of three precision mechanical stopwatches and brass calipers arranged beside a burnt-orange vernier dial.

With three distinct hosted decision APIs now in general availability, Parallel published a side-by-side technical comparison on November 16 contrasting TypeSafe’s Jev 1.14, OpenAI’s Decisions API on GPT-6 Luna, and Parallel’s own web-grounded Decisions API.

Audio briefing: Comparing OpenAI Decisions API, TypeSafe Jev, and Parallel’s web-grounded Decisions API.

Pricing, limits, and API contracts

CapabilityTypeSafe Jev 1.14OpenAI Decisions APIParallel Decisions API
List pricing$0.042 / 1M input (output free)$0.10 / 1M fresh ($0.01 cached)$1 per 1,000 requests (search incl.)
Questions / callContext-bound (64K / 262K DZ)Up to 400Up to 100
Primitivesnoul, choice, score, boxpredicate, choice, scorenoul, choice, score + citations
JevBench (231 q)88.3%90.0%89.6%
PWS-Decisions (live web)33.4%41.7%76.2%
AA Decision Index (Nov 5)57.1 (#10 of 23)48.5 (#21 of 23)Not listed
Specifications and published benchmark figures compiled from vendor announcements and the November 5 Artificial Analysis snapshot.

For state-only classification where all necessary evidence already sits inside the request, TypeSafe Jev undercuts OpenAI on fresh input tokens ($0.042 vs $0.10 per million) while offering up to 255 options per choice and an experimental box primitive. OpenAI counters with prompt caching ($0.01/M cached tokens), native JSON inputs, GPT-Live integration, and 400-question batches. Parallel solves a different problem: searching the live web and returning shared citations when evidence is not in your database.

Mark Kashef walks through a practical head-to-head comparison of OpenAI’s Decisions API and TypeSafe Jev.

Reference brief (embedded PDF)

Caveats on interpreting public decision-model benchmarks

Parallel builds one of the three APIs compared here, and explicitly notes that multiple independent projects use the “JevBench” name. Third-party tracking such as the Artificial Analysis Decision Index averages five broader evaluations including hallucination resistance and tool routing, which explains why rankings diverge between suites.

Keep what you save

Sign in to keep bookmarks and private notes on every device. We email you a link; there is no password, and the same link creates your account.