Harvey Review, Pricing & Alternatives (2026)

Harvey is an enterprise legal software company founded in 2022 by Winston Weinberg and Gabe Pereyra, according to TechCrunch’s account of its origin. Its March 2026 announcement said more than 100,000 lawyers across 1,300 organizations used Harvey, while the homepage later reported 3,500-plus legal organizations, 200,000-plus professionals, and use in more than 85 countries. In September 2026, Harvey announced $550 million in funding at a $15.5 billion valuation. The company’s March announcement said it had raised more than $1 billion in total funding. Harvey remains an active legal software company. Harvey targets law firms and corporate legal departments through an enterprise sales process. Its products combine legal assistance with shared document workspaces, rather than offering a consumer-facing legal-advice subscription.

Harvey organizes work across agents, Spaces, Vault, contract intelligence, research, and Command Center. A typical engagement starts with a user uploading documents or working in a connected workspace, then asking Harvey to analyze, compare, summarize, draft, or extract information. Vault is designed for querying large document collections, while agents route tasks across complex legal work. Shared Spaces bring firms, clients, and co-counsel into one workspace with governance and ethical-wall controls. Command Center gives administrators visibility into usage across teams, practice groups, and offices. Teams can keep reference documents in Vault and use workflows to repeat a structured task. The platform also offers a Word add-in, reducing the need to move contract drafting entirely into a separate interface.

Harvey says it uses multiple foundation-model providers and lets administrators control which models are available. Its March 2026 multi-model announcement names Anthropic, Google DeepMind, and OpenAI, and discusses Claude Opus 4.6, Sonnet 4.6, Gemini 3 Flash, GPT-5.2, and GPT-5.4. Harvey announced GPT-5.4 availability in March 2026, reporting a 91.0% score on its BigLaw Bench evaluation versus 89.8% for GPT-5.2. The company also announced its first post-trained open-weight model and the Harvey LAB legal-agent benchmark in September 2026. These model and benchmark claims come from Harvey, not independent testing. Those scores are vendor-reported results on Harvey's own evaluation suite. They should guide a trial design rather than substitute for checking citations, contractual language, and the team's own representative documents before adoption.

Harvey does not publish standard seat, usage, or enterprise prices. Paid access is Custom, so buyers need a quotation covering the commercial terms for their deployment. The evaluation terms describe an authorized free trial and say Harvey may limit the resources and features available to trial users. The same terms identify the service as a research tool whose output may contain errors, misstatements, or omissions. Harvey publishes no permanent free plan. The practical cost question is therefore less about a public monthly rate and more about negotiated scope, including users, document volume, storage, integrations, support, retention, and regional processing. The evaluation terms allow Harvey to restrict trial features and computing resources. A procurement team therefore needs a written agreement covering the intended deployment instead of treating trial access as a guaranteed production allowance.

The company's security page says customer data is not used to train models. Buyers still need to review retention, access controls, and deployment terms against their own confidentiality obligations before putting client documents into production. Harvey’s clearest advantages are its breadth of legal workspaces, model selection, custom agents, document collections, and administrative controls. Its main tradeoffs are opaque pricing, enterprise deployment requirements, reliance on external model providers, and the need for lawyers to check generated work. Clio Duo, Spellbook, LexisNexis Protégé, CoCounsel, Luminance, and Ironclad AI overlap with parts of Harvey’s legal-assistance category, although their product scopes differ. Harvey is the stronger fit for larger firms and in-house departments that need shared legal workspaces, custom agents, regional processing, and governance. Smaller teams should first confirm pricing, implementation effort, and required controls.