What does JevBench show for Jev and Plumb-4B?

In v1.4.2.2, Jev is rank 4 with a JevBench Score of 63.3; Plumb-4B (crh225, JevK5 v0.2 + LoRA) is rank 2 with a score of 65.8. The table shows their published axes and sealed-set accuracy separately.

Which system has higher sealed-set accuracy?

Plumb-4B (crh225, JevK5 v0.2 + LoRA) has the higher published sealed-set accuracy (38.0% versus 36.7%). This is separate from the composite JevBench Score.

Can I compare the cost values as actual bills?

Jev 1.13.0 (TypeSafe AI): measured at $0.040 per 1,000 decisions. Plumb-4B (crh225, JevK5 v0.2 + LoRA): estimated at $0.030 per 1,000 decisions. Estimated and announced bases are not measured charges; inspect the board’s full row disclosure.

Is Plumb-4B open source, and is Jev?

The published row lists Plumb-4B (crh225, JevK5 v0.2 + LoRA) with license note “Apache-2.0 (weights and code; NOTICE credits JevK5, Qwen and SemIf); base Qwen3.5-4B Apache-2.0”. Jev 1.13.0 is listed as “proprietary API”: TypeSafe AI serves it through its own API and has not published its weights. Check each linked source for the exact terms.

JevBench by Benchmark Heaven · released v1.4.2.2

Jev vs Plumb-4B: published benchmark comparison

A side-by-side view of Jev 1.13.0 and Plumb-4B (crh225, JevK5 v0.2 + LoRA) from the same hash-checked v1.4.2.2 aggregate. The overall score is a composite; compare the separate measures against your use case.

Four-radar comparison: Jev and Plumb-4B

Four radars compare this fixed pair across the score axes, accuracy per tier, and accuracy by family on the hard tier and sealed set. Further out is better on every spoke.

A: Jev 1.13.0 — Jev (TypeSafe, closed) · Score 63.3 (#4)
B: Plumb-4B — Jev rebuild · Score 65.8 (#2)

The four score axes

0–100, the values in the table. A label-only system has no calibration (counted as 0).

Accuracy per tier, incl. sealed

Share correct per tier; Sealed = the 308 private decisions, aggregate only.

Hard tier by family (v1.2 topics)

Plumb-4B has no published v1.2 hard-tier family breakdown.

Share correct within each family of the 220 v1.2 hard-tier decisions (public and held-out).

Sealed set by family

Share correct within each sealed family — system-level aggregates; the items stay private.

Published values

Plumb-4B (crh225, JevK5 v0.2 + LoRA) has the higher published JevBench Score. The displayed score uses one decimal place; the comparison above uses the same published release.

All values as a table

Measure	Jev 1.13.0 (TypeSafe AI)	Plumb-4B (crh225, JevK5 v0.2 + LoRA)
Published rank	#4	#2
JevBench Score	63.3	65.8
Sealed-set accuracy	36.7%	38.0%
Intelligence axis	53.1	53.0
Calibration axis	76.3	75.5
Speed axis	83.3	93.5
Cost axis	52.0	55.8
Cost evidence	measured	estimated
Cost per 1,000 decisions	$0.040 per 1,000 decisions	$0.030 per 1,000 decisions

Intelligence, Calibration, Speed and Cost — equal-weight harmonic mean, with generalization and Jev-class gates. Cost evidence is labeled per row. Speed is a benchmark axis; deployment latency depends on the endpoint and conditions.

Jev 1.13.0 (TypeSafe AI)

Marked closed in the published row. License note: proprietary API.

Published source

Measured on: production API (api.typesafe.ai)
Speed adjustment: none (production API)
Cost basis: public tariff x measured tokens (https://docs.typesafe.ai/models (output tokens not billed)) [corrected in v1.2.3: the price now averages each of the 314 v1.1 decisions once; see results/v1.2/cost-correction-v1.2.3.json] | public tariff x measured tokens (hard-tier run)

Plumb-4B (crh225, JevK5 v0.2 + LoRA)

Code and weights marked open in the published row. License note: Apache-2.0 (weights and code; NOTICE credits JevK5, Qwen and SemIf); base Qwen3.5-4B Apache-2.0.

Published source

Measured on: H100 80 GB (evaluator-owned Lium pod) · our evaluator-owned Lium GPU pod (H100 80 GB (evaluator-owned Lium pod)), offline read-only container, author's model served on loopback
Speed adjustment: x2 + 0.15 s (assumption, not measured)
Cost basis: ESTIMATE: EmpirioLabs qwen3-5-4b public pay-as-you-go list price $0.04/M input, $0.07/M output; exact Qwen3.5-4B base model, 25 Sep 2026 cutoff; no output generated by this system. https://empiriolabs.ai/models/qwen3-5-4b

Compare more JevBench pairs: Jev vs Imajev-4B · Jev vs decider-4b v2 · Jev vs JevK5 · Jev vs Cygnet · Jev vs Hopper · Jev vs Winnow-12B Q8 · Jev vs reflex 4B · Jev vs Laya

See Jev alternatives · Choose by use case

Frequently asked questions

What does JevBench show for Jev and Plumb-4B?: In v1.4.2.2, Jev is rank 4 with a JevBench Score of 63.3; Plumb-4B (crh225, JevK5 v0.2 + LoRA) is rank 2 with a score of 65.8. The table shows their published axes and sealed-set accuracy separately.
Which system has higher sealed-set accuracy?: Plumb-4B (crh225, JevK5 v0.2 + LoRA) has the higher published sealed-set accuracy (38.0% versus 36.7%). This is separate from the composite JevBench Score.
Can I compare the cost values as actual bills?: Jev 1.13.0 (TypeSafe AI): measured at $0.040 per 1,000 decisions. Plumb-4B (crh225, JevK5 v0.2 + LoRA): estimated at $0.030 per 1,000 decisions. Estimated and announced bases are not measured charges; inspect the board’s full row disclosure.
Is Plumb-4B open source, and is Jev?: The published row lists Plumb-4B (crh225, JevK5 v0.2 + LoRA) with license note “Apache-2.0 (weights and code; NOTICE credits JevK5, Qwen and SemIf); base Qwen3.5-4B Apache-2.0”. Jev 1.13.0 is listed as “proprietary API”: TypeSafe AI serves it through its own API and has not published its weights. Check each linked source for the exact terms.