LegalBench-RU - Knowledge: leaderboard

The knowledge track of LegalBench-RU — 311 cases testing recall of the correct Russian statute or norm (citation and rule-recall); accuracy %.

Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 6 models tracked.

Top models

#ModelScoreOverall rank
1DeepSeek V3.253#205
2Qwen 3 235B A22B48#313
3GPT-OSS-120B21#341

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=legalbench-ru-knowledge · How It Works · Data refreshed daily, snapshot 2026-10-20.