Decision model

Jev 1.13.0

1 question per request

57.5% of the 2,198 questions right (55.4% to 59.6%, 95% interval). Back to the leaderboard.

How we called it

Kind

Decision model

Access

TypeSafe System One API

Paid as

Provider service

Model ID

jev-1.13.0

Confidence from

Option score

Run

49, 4 October 2026

Settings

Reasoning

off

Request format

1 question per request

Requests at once

1

Temperature

not recorded

Harness seed

0

Harness commit

not recorded

Prompt version

not recorded

Results

Comparison group: xyzAB. Compare with another setup.

Confidence

Answers kept at 99% or more

312, of which 1 wrong (share right: 98.2% to 99.9%, 95% interval); 1,886 skipped, the model less than 99% sure

Calibration gap

5.1 points

AUROC, all questions

0.81 0.79 to 0.83

AUROC, reading

0.89 0.86 to 0.91

AUROC, grammar

0.72 0.70 to 0.75

Brier score

0.177

Accuracy of its surest answers

most confident fifth: 98.2% right; most confident half: 81.4% right; most confident four fifths: 64.7% right

Typical score of a right answer

80.0%

Typical score of a wrong answer

43.0%

Repeats

Run 5 times on all questions.

Answers changed in 5 runs

314 of 2,198

Same wrong answer in all 5

721 of 2,198

Wrong in at least one run

1,035 of 2,198

Accuracy in each run

57.5%, 57.1%, 57.6%, 57.3%, 58.0%

Latency

Latency, median request

0.243 s (its repeated runs)

Latency, 90th percentile

0.281 s

Requests at once

1

Throughput

251.7 questions per minute of request time

Request time

8.8 minutes, over the requests whose answers were kept

Latency: the time from sending one request to receiving its answer, measured from our computer, network and any retries within the request included. Request time: the time at least one request was in flight, over the requests whose answers were kept. It is not the run's elapsed time: it leaves out requests that failed and any pause before a run was resumed.

Cost

$ per 1,000 questions

$0.029

Whole run

$0.064

Basis

Estimated from published prices and recorded token usage

Price

$0.042 input, output free per 1M tokens (source, 2026-10-08)

Input tokens per question

687

Output tokens per question

52

Going first in front of an LLM

Replay of recorded answers, cutoff chosen on half the questions, averaged over the splits. Savings are estimates from API prices; the Method page explains how.