German Artificial Analytics
a PeerBench project

German LLM Benchmark · Model Profile

Fable 5

Anthropic provider-internal run 2026-07-06
80 .2%

avg. German score

#5 of 31 models

+7.5pp above avg.

Benchmark breakdown

GermEval

87.7%

Native German named-entity recognition — identify persons, locations, organisations and misc entities in German text, emitted as JSON. Scored with seqeval micro-F1 excluding the noisy MISC class. Run reasoning-off.

Named-entity recognition native · Native German
via GermEval (via EuroEval) ↗

INCLUDE

72.7%

Native German exam and licensing questions covering region-specific knowledge — history, law, civics and culture. Written by humans in German, not translated.

4-option multiple choice native · Native German
via CohereLabs/include-base-44 ↗

MuSR

89.5%

Multi-step soft reasoning over long narrative contexts — murder mysteries, object placement and team allocation. Requires chaining clues across several paragraphs to reach the correct answer. Translated to German from the original English MuSR benchmark.

2–5 option multiple choice translated · Professional translation
via zayne-sprague/MuSR ↗

SB10K

67.3%

Native German social-media sentiment classification — positive, neutral or negative. Human-annotated German text, not translated. Run reasoning-off; headline score is MCC on the predicted label.

3-class sentiment native · Native German
via SB10K (via EuroEval) ↗

ScaLA

83.6%

Native German linguistic acceptability — does the sentence read as grammatical German (ja / nein)? Built from clean vs. minimally-corrupted German sentences. Run reasoning-off.

Binary acceptability native · Native German
via ScaLA-de (via EuroEval) ↗

Cost & speed

$31.448 per 1,000 questions