GLiNER2 large (Fastino)

Data: JevBench v1.6.1, published 2026-10-06

Capability 32.4 · Not ranked in the current release

Last measured: 2026-09-28 · measurement release: v1.5.0

Apache-2.0

Base model: microsoft/deberta-v3-largesource

intelligence

22.5

calibration

42.4

speed

64.8

cost

77.6

Composite (secondary)

8.4 · Not ranked on this board

Cost and measurement conditions

$0.0056 per 1,000 decisions (estimated)

ESTIMATE: v1.5 measured-input proxy; no exact base-model floor applies

p50 latency: 1.9 s

x2 + 0.15 s (assumption, not measured)

evaluator-owned CPU container, offline read-only, pinned cores

Published source