Updated Oct 2026: competitor pricing figures and Claude plan details re-checked against the vendors' own pages; third-party price estimates and our cost models removed.
Want the straight side-by-side first? See Harvey vs Legora, compared feature by feature, then come back for the decision most firms are actually making.
Harvey does not publish standard pricing; ask what is available to a firm of your size. Legora does not publish standard pricing either. Most US law firms are small, and for them the legal AI choice the press writes about, Harvey against Legora against Claude, is not the choice on the table.
Every other firm faces a different conversation. This piece is that conversation.
A note before we start. Where we cite a dollar figure on an Anthropic plan, the source is Anthropic's own pricing page and help center. We do not repeat third-party estimates of what Harvey or Legora charge; neither vendor publishes a standard price, and a figure that circulates online is not a price.
Three things changed the conversation in 2026. Anthropic's published Claude Enterprise pricing is a per-seat fee plus usage billed at API rates.6 A federal judge issued the first decision on AI and attorney-client privilege in February.1 And Claude Cowork's memory is personal to the user, not a managed firm record. Together these reshape the legal AI decision into a sequence of questions a careful managing partner has to answer in order. Each one changes the calculus on the next.
Pro and Max have audit, retention, and contractual gaps that no user toggle fixes.
The relevant ruling is United States v. Heppner, S.D.N.Y., February 17, 2026, Rakoff, J.1 The defendant in Heppner used Claude on his own, without counsel's direction, to draft his defense strategy. The court held those communications were not privileged: "Because Claude is not an attorney, that alone disposes of Heppner's claim of privilege."
The holding is narrower than the headlines made it sound. Judge Rakoff noted in dictum that attorney-directed use might be analyzed differently as a Kovel-style agent. So Heppner did not categorically ban attorney use of Claude. It is unsettled doctrine.
The harder problems for Pro and Max are structural and exist regardless of how Heppner is interpreted. There is no Zero-Data-Retention addendum available on Pro or Max. There is no SSO, no SCIM, no administrator audit logs, no BAA.[3,4] Pro and Max keep a chat history, but they have no administrator audit logs. If a privilege challenge arises later, there is no firm-controlled record to demonstrate that the use was carefully structured.
Anthropic's Consumer Terms govern the Pro and Max relationship regardless of which user toggles are flipped. Anthropic states that those terms "do not apply to services under Commercial Terms, including Claude for Work..."2 Turning off training does not move a Pro or Max user onto Commercial Terms. The contractual data-handling regime is set by the plan, not by user-level switches.
What that means in practice for any attorney trying to use Pro or Max for client-confidential work. Every input is a potential privilege-waiver risk. Each prompt would need to be vetted in real time against waiver doctrine. There is no administrator audit log to show how each input was scoped. There is no contractual ZDR commitment. Consumer Terms are the operating contract regardless of personal settings. Attorneys should not have to second-guess every keystroke while they are trying to do their work.
A reasonable attorney can decide whether those gaps are tolerable for their practice. Most firms doing serious risk review conclude they are not.
Harvey and Legora do not publish a price.
For solo lawyers, boutiques, and mid-market firms, the first question is whether either incumbent will quote at all, and on what terms; neither publishes standard pricing, so the only way to know is to ask. The realistic 2026 choice for most of those firms is not Harvey vs. Legora vs. Claude. It is Claude vs. a legal-native platform built for the way lawyers actually work.
For a larger firm, both may be reachable. The question is then where partner-level money goes: a sales-led platform, or infrastructure you control. The rest of this piece is about the second option, because it is the one most firms underestimate.
Cowork memory is personal, not a firm record. The lawyer becomes the operating system.
This is the question most firms underestimate, because it does not show up in the procurement spreadsheet.
Claude Cowork's memory is personal to the user, not a managed firm record; confirm what it retains and who can see it. To get firm-level continuity, the lawyer has to save key files into project folders, tell Cowork where to look at the start of a new session, and ask it to re-read those files before new work begins. Cowork re-tokenizes that context every time and bills for it. The model does not know what the firm decided about this matter last week. It does not know the firm's standard positions. When the lawyer leaves, the project structure and the institutional knowledge of how to make Cowork useful walk out the door with them.
This is an architectural property of foundation-model products, not a defect to be patched. A foundation model is a stateless reasoning engine. Persistent matter context, firm memory, accumulated playbooks, and team-shared institutional knowledge are capabilities a firm has to build on top of it. The lawyer ends up doing that work in non-billable time, every day.
The seat fee is not the price.
Anthropic's published Claude Enterprise pricing is a per-seat fee plus usage at API rates.6 Anthropic's help center puts it plainly: the seat fee covers access, and usage is billed separately.4 So cost depends on use.
The one usage figure Anthropic publishes is for Claude Code: an enterprise average near $13 per developer per active day, or roughly $150 to $250 per developer per month.5 That number comes from software engineers. Document-heavy legal tasks can use many tokens, since a contract review or a citation-grounded research question carries the documents themselves into every turn. We do not publish our own token model for legal work; ask Anthropic for a usage estimate for your matters, and read the seat fee as the floor, not the price.
The build burden absorbs the headline savings.
Claude's published materials do not describe a built-in legal citator. Legal-specific skills and playbooks, citation checking, matter-centric context and firm memory, jurisdictional research grounding, audit trails, onboarding, governance, integration build and ongoing maintenance against firm systems, skill versioning across every model release, and a Word integration where lawyers actually live: all of it sits above the model.
What a firm running its own Claude stack takes on, as a list rather than a budget:
- Legal engineering and skill maintenance. Someone has to own the prompts, playbooks and skills, and keep them working.
- Integration maintenance. iManage or NetDocuments, Westlaw or Lexis, Outlook, billing, e-sign, conflicts. Each integration is a living dependency.
- Skill and prompt versioning across model releases. Each release can need regression testing, re-prompting and edge-case handling.
- Citation, audit, and jurisdictional tooling. Licences for citation checking, audit logging beyond the Claude Enterprise baseline, and jurisdictional research.
- Onboarding, change management, IT governance. Training, plus someone who owns token-budget alerts and model-selection policy.
Raw Claude looks like a large saving on paper. Once the build is loaded in, the comparison narrows, and how far it narrows depends on figures only your firm can put a number on.
Every Anthropic model release becomes an unscheduled engineering project.
Anthropic ships new models often, and each one can require re-testing custom prompts: regression-testing every custom skill, re-tuning prompts that drift, handling new edge cases, re-running integration tests, and explaining to partners why yesterday's workflow is producing different output today. The firm did not choose any of these projects. They show up on the calendar whether or not the firm has the engineering capacity for them.
A vendor can spread this maintenance across its customer base. A single firm absorbs it alone. Integrations and skills are not set-it-and-forget-it infrastructure. They are living dependencies that change with every upstream release.
A firm choosing to run its own Claude stack is choosing to inherit a delivery calendar set by Anthropic, not by the firm.
The decision your firm is actually making.
For a solo or small firm. Harvey and Legora do not publish pricing for a firm your size; ask, but do not plan around it. Pro and Max have the audit, retention, and contractual gaps described above. Claude Team starts at two seats and includes SSO; audit logs, zero data retention and a BAA are Enterprise features.3 Your real decision is Claude Team plus the work you add on top, vs. a legal-native platform built for your actual practice.
For a mid-market firm. Harvey and Legora may quote; the price is theirs to state, not ours. Claude on its own carries the build and maintenance list above, continuing indefinitely. Your real decision is whether to absorb the engineering and the maintenance calendar, or buy a platform that already includes them.
For an Am Law 100. Both incumbents are within reach and sales-led. Claude on its own scales with usage and with the size of the team maintaining it. Your real decision is whether a sales-led platform or the build burden is the better use of partner-level money.
The underlying question in every case is the same. What should attorneys be spending their time on? The right answer is the practice of law. Not maintenance of a Claude integration. Not real-time vetting of every prompt for waiver risk. Not manual re-uploading of context Cowork did not keep.
Irys includes legal workflows, matter context and a Word add-in, so a firm does not build them itself, with no per-token or per-query overage on the advertised Irys One plan. The real decision is whether to keep paying attention to the press's three-way comparison, or to make the comparison that actually applies to your firm.
Sources
- [1] United States v. Heppner, S.D.N.Y. (Rakoff, J., Feb. 17, 2026). Discussed in Harvard Law Review (Mar. 2026), harvardlawreview.org/blog/2026/03/united-states-v-heppner.
- [2] Anthropic, Updates to Consumer Terms and Privacy Policy. anthropic.com/news/updates-to-our-consumer-terms
- [3] Anthropic Help Center, What is the Team plan? support.claude.com/.../what-is-the-team-plan
- [4] Anthropic Help Center, What is the Enterprise plan? and How am I billed for my Enterprise plan? support.claude.com/.../what-is-the-enterprise-plan
- [5] Anthropic Claude Code Docs, Manage costs effectively. code.claude.com/docs/en/costs
- [6] Anthropic, Claude pricing, Enterprise: seat price plus usage at API rates. claude.com/pricing
---
Go deeper on the platforms discussed here: an independent, sourced Harvey vs Legora comparison, the same treatment for Harvey vs CoCounsel, and the Harvey AI pricing guide separating what is confirmed about Harvey's cost from what circulates unverified.
