Harvey on GPT 5.5, Clio vs the Status Quo, Legal Innovators +

· Source: Artificial Lawyer · Field: Legal & Regulatory — Legal Technology (LegalTech), Corporate Law & Business Legal Services, Compliance & Risk Management · Depth: Intermediate, medium

Summary

Harvey's BigLaw Bench (BLB) evaluation suite assessed OpenAI's new GPT-5.5 frontier model, reporting a score of 91.7%, an increase from GPT-5.4's 91.0%. The model demonstrated improved substantive accuracy, stronger organizational structure, and consistent formatting across legal practice areas, with particular strength in risk assessment, deal management, and litigation filing analysis. It achieved 43% perfect scores and 87% of tasks above 0.80. Anthropic's Claude Opus 4.7 also showed a 0.7% improvement over its predecessor, 4.6. While the overall percentage gains for both models are smaller than previous iterations, industry leaders like Clio's Ed Walters and Jack Newton emphasize the critical need for legal firms to adopt AI, asserting that clients now expect AI integration and that firms not embracing it risk falling behind.

Key takeaway

For CTOs and VPs of Engineering in legal firms, the incremental but consistent gains in models like GPT-5.5 and Claude Opus 4.7, coupled with strong client and talent expectations, necessitate a proactive AI adoption strategy. Your firm should prioritize integrating AI tools to enhance substantive accuracy and operational efficiency, or risk becoming uncompetitive as clients increasingly demand AI-driven legal services and young lawyers seek AI-forward workplaces.

Key insights

AI model improvements, though incremental, are driving significant shifts in legal practice expectations and operational efficiency.

Principles

In practice

Topics

Best for: CTO, VP of Engineering/Data, Executive, Legal Professional, Director of AI/ML, Consultant

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Lawyer.