Skip to content

The Cost Reality

No Claude Code required

What AI tools actually cost, who it makes sense for, and who should start smaller.

The Three Platforms

AI tool pricing changes frequently. Verify at each provider's website for current rates.

Claude (Anthropic)

Plan Monthly Cost What You Get
Free $0 Sonnet model with daily limits. Enough to try it.
Pro $20 Full chatbot + Claude Code access. Enough for regular daily use.
Max $100-200 5-20x more usage than Pro. For heavy Claude Code users and extended projects.

Coding agent: Claude Code. A terminal-based agent that runs on your computer — reads files, writes and runs code, connects to email and calendar. Included with Pro and Max plans (no additional cost). Also available via API for pay-per-token usage. This is the primary coding agent covered in this guide.

Also: Co-work. A visual desktop automation tool — Claude operates your computer through a point-and-click interface rather than a terminal. Included with Claude subscription.

Claude is our primary recommendation. Pro ($20) is sufficient for most faculty. Max is worth it if you find yourself hitting usage limits regularly.

Verify at claude.ai/pricing.

ChatGPT (OpenAI)

Plan Monthly Cost What You Get
Free $0 Standard model with daily limits. Functional but constrained.
Plus $20 Full access to all models, Deep Research, image generation, web browsing.
Pro $100-200 Higher usage caps, access to strongest reasoning models, higher Codex limits.

Coding agent: Codex. OpenAI's coding agent, included with ChatGPT plans (no separate subscription). Runs tasks asynchronously — you assign work and it completes it in the background, similar to assigning a task to a research assistant. Integrated into the ChatGPT interface.

ChatGPT's other standout feature is Deep Research — a web-browsing tool that spends 5-30 minutes searching and synthesizing sources. It is the best current tool for literature landscape scans and checking the status of pending legislation. Worth having alongside Claude.

Verify at chatgpt.com/pricing.

Gemini (Google)

Plan Monthly Cost What You Get
Free $0 Basic Gemini with daily limits.
AI Pro $20 Full access to Gemini Pro models, NotebookLM upgrades, Google Workspace integration.
AI Ultra $250 Highest-capability models, maximum usage.

Coding agents: Jules and Antigravity. Jules is a repo-aware agent you assign tasks to (included with AI Pro and Ultra at higher limits). Antigravity is a full coding environment with AI agents built in. Both are newer and less mature than Claude Code or Codex.

Gemini's biggest advantage for non-coding work is Google Workspace integration — if your institution runs on Google (email, calendar, docs), Gemini works natively with all of it. NotebookLM is also a genuinely unique tool for research synthesis.

Verify at one.google.com/about/google-ai-plans.

What to Actually Buy

The full stack (~$120/month):

Tool Monthly Cost
Claude Max ~$100
ChatGPT Plus ~$20
Total ~$120/month

That is ~$1,440/year. It is real money — but compare it to a single hour of outside counsel or a semester of a research assistant's time, and the math starts to look different.

The minimum viable setup (~$20/month):

Tool Monthly Cost
Claude Pro ~$20
Total ~$20/month

You lose ChatGPT's Deep Research and some of Claude Code's heavy-usage capacity, but you still get the core: a strong AI chatbot with structured prompting and Claude Code access. You can always upgrade as your needs grow.

Who Should Pay This

Law Faculty with Research Support

If you have a research budget, start-up funds, or a discretionary account, this is straightforward to justify. At a typical research assistant rate of $20-25/hour, the full stack costs the equivalent of about 5-6 RA hours per month. If it saves you more than that in research and writing time — and it will, once configured — the math works.

Check with your institution first. Many law schools are negotiating enterprise agreements with AI providers. Some already cover Claude or ChatGPT subscriptions for faculty. Your dean's office, IT department, or faculty services coordinator can tell you what is available before you pay out of pocket.

For lawyers billing at professional rates, the productivity gains from faster research, drafting, and document analysis justify the cost immediately. If an AI-assisted memo saves you two hours per month, the subscription pays for itself many times over.

Who Should Start Smaller

Adjunct faculty, fellows, and those without institutional support face a real trade-off. $1,440/year is not trivial if it comes entirely out of pocket. Start with Claude Pro at $20/month and add tools as you see the value.

ROI Framework

The research-time calculation: Think about the tasks that consume your time but do not require your unique expertise — initial case research, first-draft memos, formatting citations, summarizing articles for a literature review, generating exam hypotheticals. If AI handles the first 70% of those tasks and you spend your time on the last 30% (the part that requires judgment), how many hours per month does that free up?

Conservative estimate: If this system saves you 5 hours per month of routine research and drafting time, and your time is worth $50/hour (a modest estimate for a law professor's opportunity cost), that is $250/month in recovered time. Net gain: $130/month.

Realistic estimate with a configured system: Once you have project folders, a voice file, and practiced prompting habits, 10-15 hours/month of time savings is achievable. That is $500-750/month in recovered time — time you can redirect to scholarship, teaching preparation, or simply leaving the office at a reasonable hour.

The honest caveat: There is an upfront time investment to learn the tools and build the system. We estimate 8-12 hours over the first two weeks to work through the essentials and build your first project folders. The ROI is real but not instant.

Data policies differ by product

Claude Code uses the Anthropic API — your conversations are not used for model training under the current API data policy. Claude.ai (the browser version) is a consumer product with a different policy — check your account settings for data usage preferences. If you work with confidential client information or sensitive research data, this distinction matters. See Privacy for more.

What to Try First

If you are not sure whether this is worth it, here is a low-risk sequence:

  1. Week 1: Sign up for Claude Pro ($20/mo). Work through Chatbots Done Right and Prompt Engineering. Use Claude for one real work task daily — a case summary, a memo outline, a discussion prompt for class.
  2. Week 2: Build your first project folder for a recurring task. Build your voice file. Notice whether the AI's output is noticeably better with context.
  3. Week 3: Try ChatGPT Plus ($20/mo) alongside Claude. Use ChatGPT's Deep Research for a literature review or regulatory landscape scan. Compare the outputs.
  4. Week 4: If you want to go deeper, install Claude Code (Install Guide) and try the Setup path.

At each step, you can stop. You do not need the full stack to get value. The chatbot alone, used well, is worth the subscription for most law faculty.

Getting started

Working through this guide from the beginning, most faculty have a working AI setup — project folders, voice file, practiced prompting habits — within a few hours. The techniques work regardless of which subscription tier you choose.