Benchmark Almanac

BenGER

Gemini 3.1 Pro 71.1 leads since May 2026
Not in model cards

Models write German legal analyses for 596 exam-style cases and 531 doctrinal reasoning tasks. A rubric judge scores legal conclusions, reasoning, and presentation against expert references. Case scores map to the German 0-to-18 grading scale, with four points required to pass.

The August 2026 revision rescores answers with GPT-5.4-mini and compares six judge families. Human calibration uses a controlled subset. Generation counts differ across models, and model scores do not establish reliable legal advice.

Released
May 2026
Kind
Law
Level
Expert
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Subject
Law
Language
German
In and out
Text → Text
Interaction
Single turn
Items from
Written by experts
Refresh
One fixed set
Run by
Anyone
In views
Law and finance
Strongest model tested
GPT-5.4 #28 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026

In the authors’ words

We introduce BenGER (Benchmark for German Law), a benchmark and dataset for evaluating LLM systems on subsumption-based legal reasoning in German law. The dataset combines 596 exam-style free-text legal case tasks across multiple levels of legal education and 531 short doctrinal reasoning tasks. It includes a controlled validation subset of timed human-written solutions under both unaided and human-AI co-creation conditions. We evaluate 12 contemporary LLM systems - closed flagship, efficiency-oriented, and open-weight - with a rubric-aligned LLM-as-a-Judge cross-validated against a multi-rater human-grading layer (three blind reviews per solution, six judge families benchmarked against the human pool). Closed-flagship systems lead the leaderboard across all three corpora, human-AI co-creation measurably improves on unaided human work, and the LLM judge tracks human grading at Pearson r=0.76 and Cohen's k=0.60. System rankings are stable across judge families and two judges from independent providers clear the Calderon single-reviewer replacement bar on human-authored solutions.

Abstract of BenGER: Benchmarking LLM Systems on Subsumption-Based Legal Reasoning in German Law (arXiv:2605.28183, CC0)
12
Models scored on its own leaderboard

Gemini 3.1 Pro leads with 71.1 among models released by April 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Gemini 3.1 Pro 71.1
  2. 2 Claude Opus 4.7 70.4
  3. 3 GPT-5.4 69.1
  4. 4 Claude Sonnet 4.6 64.0
  5. 5 Gemini 3.1 Flash-Lite 60.5
  6. 6 GPT-5.4 Mini 60.0
  7. 7 DeepSeek-V4-Pro 59.9
  8. 8 DeepSeek-V4-Flash 55.3
  9. 9 Qwen3.5-122B-A10B 54.5
  10. 10 Qwen 3.6 35B-A3B 50.6
  11. 11 Qwen3-235B-A22B-Thinking Jul 2025 48.4
  12. 12 Llama 4 Maverick 48.0

Source: Official BenGER leaderboard

#ModelScoreSettingDateSource
1 Gemini 3.1 Pro 71.1 Feb 2026 Official leaderboard
2 Claude Opus 4.7 70.4 Apr 2026 Official leaderboard
3 GPT-5.4 69.1 Mar 2026 Official leaderboard
4 Claude Sonnet 4.6 64.0 Feb 2026 Official leaderboard
5 Gemini 3.1 Flash-Lite 60.5 Mar 2026 Official leaderboard
6 GPT-5.4 Mini 60.0 Mar 2026 Official leaderboard
7 DeepSeek-V4-Pro 59.9 Apr 2026 Official leaderboard
8 DeepSeek-V4-Flash 55.3 Apr 2026 Official leaderboard
9 Qwen3.5-122B-A10B 54.5 Feb 2026 Official leaderboard
10 Qwen 3.6 35B-A3B 50.6 Apr 2026 Official leaderboard
11 Qwen3-235B-A22B-Thinking Jul 2025 48.4 Jul 2025 Official leaderboard
12 Llama 4 Maverick 48.0 Apr 2025 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: Benchathon, ZJS and Doctrinal Principles. Each ranks models on one part of the test. Gemini 3.1 Pro leads 2 of them, and Claude Opus 4.7 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Benchathon

12 models we track have a score here.

  1. 1 Claude Opus 4.7 ↑1 69.3
  2. 2 Gemini 3.1 Pro ↓1 68.3
  3. 3 GPT-5.4 66.6
  4. 4 Gemini 3.1 Flash-Lite ↑1 63.5
  5. 5 Claude Sonnet 4.6 ↓1 58.9
  6. 6 GPT-5.4 Mini 58.6
  7. 7 DeepSeek-V4-Pro 56.1
  8. 8 Qwen3.5-122B-A10B ↑1 49.4
  9. 9 DeepSeek-V4-Flash ↓1 49.3
  10. 10 Qwen 3.6 35B-A3B 42.4

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind
Own leaderboard Not tested since May 2026

Source: Official BenGER leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
50.0 60.0 70.0 2025 2026 50.0 60.0 70.0 2025 2026

Source: Official BenGER leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for DeepSeek-V4-Flash. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (12 entries).
  • Pass Models scored 12 models we track.
  • Warning Newest score The newest came 7 months ago, in April 2026, for DeepSeek-V4-Flash.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 (#28).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 12 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check