Qwen3.7 Max leads LLM Stats legal rankings

Qwen3.7 Max leads the LLM Stats legal category with a score of 44.5, followed by Qwen3.7-Plus at 41.3 and Qwen3.6 Plus at 40.9. The ranking evaluates underlying model capabilities rather than complete legal software products. Its methodology was reviewed internally, but not by a practicing attorney.

The category combines available legal benchmarks covering knowledge, applied reasoning, document analysis and related tasks. Where a category index is available, LLM Stats uses an uncertainty-aware conservative score; otherwise, it relies on benchmark coverage and normalized performance. Scores support comparisons within the category, not estimates of the probability of correct legal advice. Missing results and narrow evaluation coverage warrant caution.

Strong benchmark performance does not establish citation validity, current-law coverage, confidentiality safeguards or professional accountability. Firms should test shortlisted models inside their intended products using representative documents and known answers, with attention to jurisdiction, source grounding, permissions, retention and review procedures. For contract analysis, testing should reflect the firm’s playbook and governing law. For high-volume review, consistency, recall, false positives and cost may matter more than a small ranking advantage.