WaterSheep (Samrat Dutta, ModernBERT-base encoder with calibrated heads, served by the author's own `watersheep --serve`)v1.7 · 10 s rule

Data: JevBench v1.8.18, published 2026-11-20 · method v1.7 (10 s rule), re-scored from the recorded per-decision times

Capability 35.5 · rank #230 of 251 on the open-weights board

Last measured: 2026-10-07 · measurement release: v1.6.1

apache-2.0

Base model: answerdotai/ModernBERT-basesource

intelligence

5.3

calibration

65.8

speed

77.4

cost

68.7

Composite (secondary)

0.2 · Composite rank #243

Cost and measurement conditions

$0.011 per 1,000 decisions (estimated)

ESTIMATE, labelled, and RUN 48'S RATE FOR THIS EXACT ROW AND BASE, CARRIED UNCHANGED: answerdotai/ModernBERT-base IS in the frozen pricing_v15.UNLISTED_BASE_MODELS, so no exact base-model floor applies and the module takes the estimate rate from the row's own registry entry; USD 0.03/M input, 0 output is the rate run 47 (quyet-10-small-en) and run 48 (watersheep on v1.5) already assigned to this base. Output 0 is literal: watersheep/infer.py returns output_tokens 0 and the serving path has no generation step (published 02 Oct rule). x this row's own measured input tokens. Estimated, not charged. No new decision.

Measured p50 latency: 0.463 s (on-CPU, no network, loopback; not directly comparable with API latency, which we time end to end from our measuring host, network included)

Ranking and cap use: 1.08 s (est. incl. network (2x + 150 ms))

Self-hosted model: measured p50 0.463 s on-CPU, no network (timed on our CPU host over loopback). Not directly comparable with API latency, which we time end to end from our measuring host, network included. The ranking and latency cap use the same estimate for every self-hosted row: 2 × measured + 150 ms = 1.08 s. Model authors can ask us to time a hosted endpoint instead (on request).

x2 + 0.15 s (assumption, not measured)

measured 2026-10-07 on the live v1.6.0/v1.6.1 pool (sha256 901983ae...), 1,500 of 1,500 answered, 0 invalid. RE-MEASUREMENT on v1.6 of the row run 48 measured on the retired v1.5 basis, same weights: samratduttaofficial/WaterSheep @ ebf3705d2d9dc648d5cc6b4cbca7646c6ab48998 (Apache-2.0), package github.com/SamratDuttaOfficial/WaterSheep @ 3ccf02b3 via the author's own install line in an isolated venv (torch 2.14.1+cpu, transformers 5.19.0), served with the author's own command `watersheep --model samratduttaofficial/WaterSheep@ebf3705d... --serve --device cpu`. Their server speaks POST /v1/systemone with Jev's shapes, so the UNCHANGED typesafe adapter applies; all eight checks of the unchanged smoke gate passed on synthetic requests before any sealed item was sent. Their own max_len 512 (watersheep.json) truncates every longer state, the 23 long items included, without refusing; not overridden. RUN ON SANDY'S SHARED CPU (12-core Zen 2, 8 threads) with server and driver inside a loopback-only network namespace and no credential in the environment; load average 8.6-15.4 (median 13.8) logged every 60 s, so the Speed axis is a DECLARED LOWER BOUND. Latency axis uses the standard self-hosted x2 + 0.15 s adjustment. Serial, one decision per request, no retry, no batching, offline. add-requests run 59, benchmarkheaven.com/submit ref 6F1561FC; previously GitHub #179

Measured latency over 1,500 decisions: p50 0.463 s · p95 0.773 s

1 decision over 10 s (1 of them answered without error, counted as wrong).

A decision that takes longer than 10 seconds (measured time) counts as a wrong answer and as zero calibration for that item: each type's calibration is multiplied by the share of its decisions answered within 10 s. The limit uses the measured time (raw). The displayed adjusted time for self-hosted rows (2 × + 0.15 s) is an assumption and is not used for the limit.

Breakdowns (not part of the score)

Chance-corrected competence (0 = chance, 100 = perfect) by category on S+P+L1+L2+L3. Raw values, not equated; the same cells feed the compare view.

Capability by subject topic

Rules, policy & law0.0n=1616
Coding & software0.0n=382
Math & numbers0.0n=370
Finance & commerce0.0n=223
Support & operations0.0n=160
Everyday language0.0n=145
Safety & security0.0n=109

Use cases (TypeSafe categories)

Model routing0.0n=524
Legal & compliance0.0n=515
Customer support0.0n=240
E-commerce0.0n=119
Insurance claims0.0n=119
Risk assessment0.0n=118
Financial crime0.0n=101
Feature extraction0.0n=99
Lead generation0.0n=95
Recruiting0.0n=94
Knowledge graphs0.0n=94
LLM guardrails0.0n=93
Moderation0.0n=93
Code linting0.0n=90
Search & retrieval0.0n=90
Science0.0n=90
Advertising0.0n=88
Gaming0.0n=86
Demand forecasting0.0n=83

Cells need at least 30 answered items; † fewer than 30 (indicative only; drawn as hollow, lighter points on the compare radars).

JevBench v1.6.43 values (method v1.6, before the 10 s rule)

Capability 35.57 · Composite 0.19 · rank #215 on the all-systems board · #171 open weights

intelligence 5.31 · calibration 65.84 · speed 77.59 · cost 68.67

v1.6 speed basis (124 items): p50 0.401 s · p95 0.838 s

These figures are kept for comparison. The current figures above are re-scored under the 10 s rule from the recorded per-decision times and replace them.

Published source