Most people do not need three AI subscriptions. The practical decision is whether a free plan is enough, which product deserves the roughly $20 monthly slot, and whether your usage genuinely justifies a $100–$200 tier.
This guide was checked against official provider pages on July 11, 2026. Prices below are public US list prices before tax unless noted. Annual billing, regional pricing, promotions, capacity rules, and available models can change, so use the linked provider page as the final checkout source.
AI subscription prices at a glance
| Product | Free entry | Main individual plan | Higher-capacity options | Best reason to consider it |
|---|---|---|---|---|
| ChatGPT | Yes | Plus: $20/mo | Pro: $100 or $200/mo | Broad general-purpose work and OpenAI tools |
| Claude | Yes | Pro: $20/mo or $17/mo annual | Max: $100 or $200/mo | Writing, analysis, and Claude Code access |
| Google AI | Yes | AI Pro: $19.99/mo | AI Ultra 5x: $100; Ultra 20x: $200 | Gemini plus Google ecosystem benefits |
| Microsoft Copilot | Yes | Microsoft 365 Premium: $19.99/mo | Business and enterprise plans | Copilot inside Microsoft 365 apps |
| Grok | Yes | SuperGrok: $30/mo | Check current checkout | Real-time X context and xAI models |
| Perplexity | Yes | Pro: $20/mo or $200/yr | Max: $200/mo or $2,000/yr | Citation-oriented research and search |
| Mistral Le Chat | Yes | Pro: $14.99/mo | Team: $24.99/user/mo | Lower-cost general assistant option |
| Cursor | Limited Hobby plan | Pro: $20/mo | Pro+: $60; Ultra: $200 | AI-first coding inside an editor |
| GitHub Copilot | Yes | Pro: $10/mo | Pro+: $39; Max: $100 | IDE-native coding across the GitHub ecosystem |
The short recommendation
- Stay free if you use AI occasionally, are still learning what matters, or can tolerate variable limits.
- Choose one $20 plan when AI is part of your weekly work. Pick based on workflow and ecosystem, not a universal leaderboard.
- Choose a coding plan separately only if coding is central to your work; GitHub Copilot and Cursor solve a different job from a general chat subscription.
- Move to $100–$200 only after measuring usage. Higher tiers mostly buy capacity, priority, or premium access. They do not automatically make every answer five or ten times more valuable.
ChatGPT pricing
OpenAI currently lists a free plan, ChatGPT Plus at $20 per month, and Pro options at $100 and $200 per month. OpenAI’s model availability and usage allowances change frequently, and some models can have separate limits even on paid plans. Treat “unlimited” marketing language as subject to the provider’s terms and safeguards.
Choose ChatGPT when you want one broad workspace for writing, research, files, images, voice, and OpenAI’s wider tool ecosystem. Avoid upgrading solely because a temporary launch promotion sounds generous; evaluate your normal weekly usage after the promotion is gone.
Sources: ChatGPT pricing, OpenAI Pro plan guidance
Claude pricing
Anthropic lists Claude Pro at $20 per month, or the equivalent of $17 per month with annual billing. Max plans are listed at $100 and $200 per month. Anthropic describes usage as variable rather than promising a fixed token allocation for every account and workload.
Claude is a strong fit for people whose priority is careful writing, long-document work, analysis, or Claude Code. The right upgrade test is simple: if Pro limits interrupt paid work often enough to cost more than the price difference, Max may be rational. Otherwise, it is excess capacity.
Sources: Claude pricing, Claude Pro usage guidance
Google Gemini and Google AI pricing
Google AI Pro remains $19.99 per month. Google now offers two Ultra capacity levels: AI Ultra 5x at $100 per month and AI Ultra 20x at $200 per month. Google describes Flow credits and storage as part of these bundles, so compare the full package—not only chatbot access.
Google’s free options are meaningful for experimentation. Google AI Studio can be used without charge in available regions within applicable model limits; paid Gemini API usage is a separate billing decision. For teams already centered on Google products, bundle value may matter more than a narrow model comparison.
Sources: Google AI plans, Google’s 2026 subscription update, Google AI Studio billing
Microsoft Copilot pricing
For individuals, Microsoft’s current consumer offer is Microsoft 365 Premium at $19.99 per month, combining Microsoft 365 benefits with Copilot capabilities. Copilot Pro is a legacy plan that some existing subscribers may retain; it should not be treated as the current default purchase.
Business pricing depends on the underlying Microsoft 365 license and whether Copilot is bundled or added separately. Because the combinations change, teams should price the complete seat—not quote the Copilot add-on without its qualifying Microsoft 365 plan.
Sources: Microsoft Copilot pricing, Microsoft 365 Premium overview
Grok pricing
xAI lists a free Grok option and SuperGrok at $30 per month. Current usage is governed by shared allowances rather than a blanket promise of unlimited use. Higher-capacity offers can change and may only be visible in the live checkout, so this guide does not repeat an unverified price for SuperGrok Heavy.
Grok’s clearest differentiator is its relationship with X and xAI’s current models. That is valuable for some real-time research workflows, but organizations with strict content governance should evaluate its controls and source quality before deployment.
Sources: xAI pricing, Grok usage guidance
Perplexity pricing
Perplexity Pro is $20 per month or $200 per year. Perplexity Max is $200 per month or $2,000 per year. The provider describes extended research, creation, and advanced-model access, but plan limits remain part of the service.
Perplexity makes the most sense when search, citations, and source discovery are central to your work. It is less compelling as a second paid assistant if your existing tool already covers your research needs reliably.
Sources: Perplexity plan comparison, Perplexity Max
Mistral Le Chat pricing
Mistral lists Le Chat Pro at $14.99 per month, a verified higher-education student plan at $5.99 per month for eligible users, and Team at $24.99 per user per month. Eligibility and duration rules apply to the student plan.
This is one of the lower public individual prices among major general assistants. It deserves consideration when cost, European provider context, or Mistral’s product ecosystem matters more than having the largest consumer feature bundle.
Source: Mistral pricing
Cursor pricing
Cursor offers a limited Hobby plan, Pro at $20 per month, Pro+ at $60, and Ultra at $200. Paid tiers include different amounts of model usage plus any bonus usage Cursor makes available. Because model costs differ, a dollar-denominated allowance is more useful than assuming a fixed number of identical requests.
Choose Cursor when repository-aware editing and agent workflows inside the code editor are the main job. A general AI subscription may complement it, but does not replace the same development workflow.
Sources: Cursor pricing, Cursor usage documentation
GitHub Copilot pricing
GitHub Copilot offers a free plan, Pro at $10 per month, Pro+ at $39, and Max at $100. GitHub moved individual plans to an AI-credit and included-usage structure in June 2026, so older “premium request” counts and a universal per-request overage price are now stale.
Copilot remains the lowest-priced mainstream paid coding option in this comparison. It is the practical first paid step for developers who want AI inside familiar IDE and GitHub workflows without moving to an AI-first editor.
Sources: GitHub Copilot plans, June 2026 billing change
Which $20 AI subscription is worth it?
There is no defensible universal winner. Use the decision below:
- Pick ChatGPT Plus for the broadest all-purpose workspace.
- Pick Claude Pro when writing, analysis, or Claude Code is central.
- Pick Google AI Pro when Google’s bundle and ecosystem are already part of your work.
- Pick Microsoft 365 Premium when Copilot inside Word, Excel, PowerPoint, Outlook, and the wider Microsoft subscription creates the value.
- Pick Perplexity Pro when research with visible sources is the primary task.
- Pick Cursor Pro when your $20 budget is specifically for software development.
If two descriptions sound equally useful, spend one week on both free tiers and record three things: tasks completed, corrections required, and interruptions caused by limits. Buy the plan that removes the most expensive friction—not the one with the longest feature list.
When a $100 or $200 plan makes sense
A high-capacity plan can be rational for a consultant, researcher, developer, or production team whose paid work is repeatedly interrupted by lower limits. It is usually poor value for light use, curiosity, or fear of missing out.
Before upgrading, estimate:
- how many work sessions hit a real limit each month;
- the billable or salary value of the time lost;
- whether a second lower-cost tool would solve the bottleneck; and
- whether the premium feature you need is actually exclusive to the higher tier.
If the measurable monthly benefit is not comfortably above the price difference, stay on the lower tier.
What changed since the April version of this guide
- Google replaced the old Ultra presentation with $100 and $200 capacity levels.
- Microsoft’s current consumer purchase shifted to Microsoft 365 Premium rather than Copilot Pro.
- GitHub introduced AI Credits and a $100 Copilot Max plan.
- Model names and fixed usage estimates across several providers became stale.
- Temporary launch promotions ended.
That pattern is the larger lesson: AI pricing pages age quickly. We now prefer stable plan prices, official sources, and explicit verification dates over brittle message-count estimates and model lists.
Bottom line
Start free. Pay about $20 only when one product has earned a permanent place in your weekly workflow. Treat $100–$200 plans as capacity purchases that need a business case. And always check the official provider link before checkout—the best pricing guide is a decision framework plus current sources, not a promise that fast-changing limits will stay fixed.