JevBench v1.3.0 · one system

reflex-27b (Qwen3.8-27B)

kshetrajna12 · MIT code; Apache-2.0 Qwen weights · open (code and weights) · jev-rebuild

JevBench v1.3.0 score

63.3

Rank #17 of 48 ranked systems.

11.1 points behind Jev 1.13.0's 74.4.

Axes

Intelligence
85.8
Calibration
86.2
Speed
67.5
Cost
32.3

Cost

$0.181 per 1,000 decisions (estimated).

github.com/kshetrajna12/reflex

See every JevBench v1.3.0 system, method and scoring on the full leaderboard.