DELTA
Models research Dutch legal questions and apply the law to facts. Lawyers judge accuracy and professional usefulness.
The project releases a public task set and a scored report.
- Made by
- Legal Benchmarks and partner law firms
- Released
- September 2026
- Kind
- Law
- Level
- Expert
- Scored
- Graded by people
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Subject
- Law
- Language
- Dutch
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- Occupations
- Legal
- In views
- Law and finance
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- September 2026
- Scores refreshed
- November 20th, 2026
Claude Fable 5.1 leads with 15.0 among models released by September 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. Every run is in the next chart.
The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- Claude Opus 5.5
- and 10 more
- 15.0 Claude Fable 5.1
- 13.3 Claude Opus 5
- 11.7 Claude Fable 5
- 10.0 Grok 4.6
- 6.70
DeepSeek-V4-Pro
- 6.70 Claude Sonnet 5
- 6.70 GPT-6 Astra
- 5.00
Gemini 3.1 Pro
- 5.00
Gemini 3.5 Flash
- 5.00
MiniMax-M3
- 5.00 GLM-5.2
- 5.00 Grok 4.5
- 5.00 GPT-5.6 Sol
- 5.00 GPT-5.6 Terra
- 5.00
Gemini 3.6 Flash
- 5.00
Gemini 3.8 Flash
- 3.30 Claude Opus 4.8
- 3.30 Kimi K3
- 3.30
Qwen 3.8 Max
- 0.00
Mistral Large 3
- 0.00
Mistral Medium 3.5
- 0.00
Qwen3.7-Max
- 0.00 GPT-5.6 Luna
- 1 Claude Fable 5.1 Adaptive Max 15.0
- 2 Claude Opus 5 Max 13.3
- 3 Claude Fable 5 Adaptive Max 11.7
- 4 Grok 4.6 Max 10.0
- 5
DeepSeek-V4-Pro Max 6.70
- 6 Claude Sonnet 5 Max 6.70
- 7 GPT-6 Astra Max 6.70
- 8
Gemini 3.1 Pro 5.00
- 9
Gemini 3.5 Flash 5.00
- 10
MiniMax-M3 5.00
- 11 GLM-5.2 Max 5.00
- 12 Grok 4.5 High 5.00
- 13 GPT-5.6 Sol Max 5.00
- 14 GPT-5.6 Terra Max 5.00
- 15
Gemini 3.6 Flash 5.00
- 16
Gemini 3.8 Flash 5.00
- 17 Claude Opus 4.8 Max 3.30
- 18 Kimi K3 3.30
- 19
Qwen 3.8 Max 3.30
- 20
Mistral Large 3 0.00
- 21
Mistral Medium 3.5 0.00
- 22
Qwen3.7-Max 0.00
- 23 GPT-5.6 Luna Max 0.00
Source: Official DELTA leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Fable 5.1 | 15.0 | Adaptive Max | Sep 2026 | Official leaderboard |
| 2 | Claude Opus 5 | 13.3 | Max | Jul 2026 | Official leaderboard |
| 3 | Claude Fable 5 | 11.7 | Adaptive Max | Jun 2026 | Official leaderboard |
| 4 | Grok 4.6 | 10.0 | Max | Aug 2026 | Official leaderboard |
| 5 | | 6.70 | Max | Apr 2026 | Official leaderboard |
| 6 | Claude Sonnet 5 | 6.70 | Max | Jun 2026 | Official leaderboard |
| 7 | GPT-6 Astra | 6.70 | Max | Sep 2026 | Official leaderboard |
| 8 | | 5.00 | Feb 2026 | Official leaderboard | |
| 9 | | 5.00 | May 2026 | Official leaderboard | |
| 10 | | 5.00 | Jun 2026 | Official leaderboard | |
| 11 | GLM-5.2 | 5.00 | Max | Jun 2026 | Official leaderboard |
| 12 | Grok 4.5 | 5.00 | High | Jul 2026 | Official leaderboard |
| 13 | GPT-5.6 Sol | 5.00 | Max | Jul 2026 | Official leaderboard |
| 14 | GPT-5.6 Terra | 5.00 | Max | Jul 2026 | Official leaderboard |
| 15 | | 5.00 | Jul 2026 | Official leaderboard | |
| 16 | | 5.00 | Sep 2026 | Official leaderboard | |
| 17 | Claude Opus 4.8 | 3.30 | Max | May 2026 | Official leaderboard |
| 18 | Kimi K3 | 3.30 | Jul 2026 | Official leaderboard | |
| 19 | | 3.30 | Aug 2026 | Official leaderboard | |
| 20 | | 0.00 | Dec 2025 | Official leaderboard | |
| 21 | | 0.00 | Apr 2026 | Official leaderboard | |
| 22 | | 0.00 | May 2026 | Official leaderboard | |
| 23 | GPT-5.6 Luna | 0.00 | Max | Jul 2026 | Official leaderboard |
The best score, 15.0, costs $1.76 a task
Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
Source: Official DELTA leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Claude Fable 5.1 | Adaptive Max | 15.0 | $1.76 | ||
| Claude Opus 5 | Max | 13.3 | $1.07 | ||
| Claude Fable 5 | Adaptive Max | 11.7 | $1.52 | ||
| Grok 4.6 | Max | 10.0 | $0.19 | ||
| Max | 6.70 | $0.16 | |||
| Claude Sonnet 5 | Max | 6.70 | $0.25 | ||
| GPT-6 Astra | Max | 6.70 | $0.67 | ||
| 5.00 | $0.22 | ||||
| 5.00 | $0.15 | ||||
| 5.00 | $0.05 | ||||
| GLM-5.2 | Max | 5.00 | $0.13 | ||
| Grok 4.5 | High | 5.00 | $0.13 | ||
| GPT-5.6 Sol | Max | 5.00 | $0.46 | ||
| GPT-5.6 Terra | Max | 5.00 | $0.26 | ||
| 5.00 | $0.16 | ||||
| 5.00 | $0.10 | ||||
| Claude Opus 4.8 | Max | 3.30 | $0.44 | ||
| Kimi K3 | 3.30 | $0.43 | |||
| 3.30 | $0.23 | ||||
| 0.00 | $0.007 | ||||
| 0.00 | $0.11 | ||||
| 0.00 | $0.10 | ||||
| GPT-5.6 Luna | Max | 0.00 | $0.12 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official DELTA leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official DELTA 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
- 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
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (23 entries).
- Pass Models scored 23 models we track.
- Pass Newest score The newest is from September 2026, for GPT-6 Astra.
- Note Frontier models Only 2 of the top 10 have a score. The strongest tested is GPT-6 Astra (#4).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on legalbenchmarks.ai: 23 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.