Realm: Legal
Realm: Legal uses multi-stage litigation, transactional and compliance tasks. A system produces a legal work product and revises it as facts, arguments, authority, jurisdiction or timeframe change.
Reported results
| Claude Opus 5 (max)Anthropic | 62.1 | |
|---|---|---|
| Claude Fable 5.1 (max)Anthropic | 60.8 | |
| Claude Fable 5Anthropic | 55.7 | |
| Kimi K3 (max)Moonshot AI | 54.1 | |
| Grok 4.6 (high)xAI | 53.8 | |
| GPT-5.6 Sol (max)OpenAI | 51.1 | |
| Muse Spark 1.3 (xhigh)Harvey | 50.3 | |
| Gemini 3.8 Flash (high)Google | 47.9 | |
| Grok 4.5 (high)xAI | 43.2 | |
| Muse Spark 1.1 (xhigh)Harvey | 42.3 |
The leaderboard includes newer models than the original three-model analysis. That earlier report found problems with selecting rules, applying facts, recognising missing information and revising later work. Those findings should not be attributed to the newer models without separate testing.
What the benchmark measures
Litigation, transactional and compliance tasks completed over several steps, using legal materials and tools as the facts or authority change.
The evaluated unit is an agent or completed task. Read the source for the exact prompt, tool and harness conditions.
How it is scored
Mean score across 35–60 weighted criteria per task, organised around issue, rule, application and conclusion (IRAC). The source also reports the best of three attempts.
Scores remain in the original unit. They are not normalised or combined with results from other benchmarks.
Sources
- Primary benchmark source
- micro1 Realm: Legal, updated 2 Sep 2026 · results 2026-09 · checked 2026-09-18
The recorded leaderboard covers 16 model configurations. Its detailed failure analysis concerns the original three models, so the IRAC breakdown should not be applied to newer entries.