German Language LLM Index
a PeerBench project

Individual experiment result

GermEval NER (DE)

Tencent HY3 Tencent

Micro-F1 excluding MISC reasoning off fp8
76.1%
Micro-F1 excluding MISC

Run details

Date
2026-08-30
Test cases
1,024
Median latency
1.68s
Total cost
Avg. prompt tokens
1,313.7
Avg. answer tokens
34.4
Avg. reasoning tokens
Quantization
fp8

About this benchmark

Native German named-entity recognition (person/location/org/misc) emitted as JSON; seqeval micro-F1 excluding MISC.

Named-entity recognition native · Native German Source ↗

Provenance

Canonical model
tencent/hy3
Run ID
2026-08-30T19-46-08__Tencent__tencent__hy3__germeval_de