Observed arrival · 2026-11-05
Lexassay’s legal-model spot check
Lexassay compares language models on fixed legal tasks and publishes their answers, scores, and citation checks.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers comparing legal AI answers and citations
- Worth noticing
- The page reports 2,644 confirmed citations and 804 invented or misattributed across its published runs.
Field notes
The results view separates published-task scores from hidden-set performance and lets readers filter by model family, cloud or local execution, score band, and cost. The page reports 12 published runs across 42 model configurations, with links to individual runs. Its methodology describes 15 fixed tasks, five hidden, and says grading uses AI agents against gold answers and primary sources. The authors characterize this as a documented spot check, not a strict benchmark.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
One card from the complete issue