Vals AI Legal Aid Benchmark Finds Models Fall Short on Local Help

NextLadder Ventures and Vals AI released the Access to Justice Benchmark showing 23 AI models failed to meet an 80% reliability threshold in providing local legal aid help across 252 scenarios. Models performed better in explaining law generally than in directing users to specific local resources. Qwen 3.8 Max, Claude Sonnet 5.5, and Wonder Full led the results but none clearly dominated.