German Language LLM Index
a PeerBench project

Individual experiment result

SB10K (DE)

DeepSeek V4 Pro DeepSeek

Matthews correlation coefficient reasoning off fp8
60.0%
Matthews correlation coefficient

Run details

Date
2026-06-09
Test cases
1,024
Median latency
1.42s
Total cost
$1.28
Avg. prompt tokens
778.6
Avg. answer tokens
2.1
Avg. reasoning tokens
Quantization
fp8

About this benchmark

Native German Twitter sentiment (positive/neutral/negative), generate_until variant; primary metric MCC.

3-class sentiment native · Native German Source ↗

Provenance

Run ID
2026-06-09T15-07-42__openrouter__deepseek__deepseek-v4-pro__sb10k_de