Muse Wonder Tops Legal Work Benchmark
AI at Meta says Muse Wonder received the highest overall quality score in Legora's November BAR legal work benchmark and performed strongly on the complex tasks required by enterprises.
View original postThe November BAR results are now available, showing how AI models compare on real-world legal tasks in Legora. A few highlights from this month’s results: → Muse Wonder ranks highest on overall quality and leads across four task types: analysis and answering, extraction, drafting, and editing. → GPT-6.2 Astra leads on comparison and review tasks, with overall quality 4% above November’s model average and cost per case 20% below average. → GPT-6.2 Sol delivers above-average quality at 78% below the average cost per case across models benchmarked in November.
View original
The November BAR results are now available, showing how AI models compare on real-world legal tasks in Legora. A few highlights from this month’s results: → Muse Wonder ranks highest on overall quality and leads across four task types: analysis and answering, extraction, drafting, and editing. → GPT-6.2 Astra leads on comparison and review tasks, with overall quality 4% above November’s model average and cost per case 20% below average. → GPT-6.2 Sol delivers above-average quality at 78% below the average cost per case across models benchmarked in November.
Body
Legal work is a tough test for AI.
Muse Wonder earned the highest overall quality score in Legora’s November BAR benchmark for legal work. A strong showing on the kinds of complex tasks businesses need help with.
AI Trends