German Language LLM Index
a PeerBench project

Individual experiment result

GermEval NER (DE)

Qwen3.8 27B (FP8) Alibaba

Micro-F1 excluding MISC reasoning off fp8
81.6%
Micro-F1 excluding MISC

Run details

Date
2026-08-15
Test cases
1,024
Median latency
3.60s
Total cost
Avg. prompt tokens
1,186.3
Avg. answer tokens
26.4
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
qwen/qwen3.8-27b-fp8
Run ID
2026-08-15T08-17-30__openai__Qwen__Qwen3.8-27B-FP8__germeval_de