Most people do not need three AI subscriptions. The useful decision is whether a free plan is enough, which product deserves one monthly slot, and whether a higher tier would remove a measurable bottleneck.
The market changed again since our previous review. Google’s lower-cost AI Plus tier deserves a place in the comparison, Grok now lists a $100 capacity tier, Claude expanded what comes with Pro, and Perplexity separated some agent work into a credit system. This guide was checked against official provider pages on August 11, 2026.
Prices below are public US list prices before tax unless noted. Annual billing, app-store charges, regional prices, promotions, usage rules, and included models can differ. Always 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 workspace with research, files, media, Codex and agents |
| Claude | Yes | Pro: $20/mo or $17/mo annual | Max 5x: $100; Max 20x: $200 | Writing, analysis, Claude Code, Cowork, Design and Science |
| Google AI | Yes | Plus: $7.99; Pro: $19.99/mo | Ultra 5x: $100; Ultra 20x: $200 | Gemini plus storage and Google product benefits |
| Microsoft Copilot | Yes | Microsoft 365 Premium: $19.99/mo | Business and enterprise plans | Copilot inside Microsoft 365 apps |
| Grok | Yes | SuperGrok: $30/mo | SuperGrok Plus: $100/mo | Real-time web and X search, media generation and xAI models |
| Perplexity | Yes | Pro: $20/mo or $200/yr | Max: $200/mo or $2,000/yr | Citation-oriented search, research and file/app creation |
| Mistral | Yes | Pro: $14.99/mo | Team: $24.99/user/mo | Lower-cost assistant plus Vibe coding tools |
| 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.
- Consider the lower-cost plans first. Google AI Plus costs $7.99, GitHub Copilot Pro costs $10, and Mistral Pro costs $14.99.
- Choose one roughly $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. The free plan now advertises unlimited text chats with GPT-5.6 Luna, subject to abuse guardrails, plus limited file uploads, image generation, deep research, Codex, and ChatGPT Work. Plus adds advanced GPT-5.6 reasoning, expanded Codex and research usage, scheduled tasks, custom GPTs, and broader ChatGPT Work access.
OpenAI’s model availability and usage allowances change frequently. The two Pro tiers primarily buy more capacity: OpenAI describes the $100 tier as 5x Plus usage and the $200 tier as 20x Plus usage, while temporary Codex promotions can differ. Treat “unlimited” language as subject to provider 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, ChatGPT Plus, OpenAI Pro tier guidance
Claude pricing
Anthropic lists Claude Pro at $20 per month, or the equivalent of $17 per month with $200 billed annually. Max offers 5x or 20x Pro usage, corresponding to $100 and $200 monthly tiers. Anthropic describes usage as variable rather than promising a fixed token allocation for every account and workload.
The $20 bundle is broader than it was: Claude Pro now includes Claude Code, Cowork, Design, Science, Research, unlimited projects, more models, and Claude for Microsoft 365. It is a strong fit when careful writing, long-document work, analysis, or coding is central. 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 now has a clearer price ladder. Google AI Plus costs $7.99 per month in the US, AI Pro remains $19.99 per month, AI Ultra 5x costs $100, and AI Ultra 20x costs $200. Google AI Plus is the notable budget addition: it offers more Gemini, Flow, Gmail, and Gemini Notebook access plus storage, without charging the full Pro price.
Pro and Ultra are bundles rather than simple chatbot upgrades. Depending on the plan and country, benefits can include higher Gemini limits, Google AI Studio, NotebookLM, Flow, Jules, Antigravity, Gemini in Google apps, storage, YouTube benefits, and other Google services. Availability and language support vary, so a benefit that appears on the US page may not be usable everywhere.
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 AI Plus launch and price, Google’s 2026 subscription update, Google AI Studio billing
Microsoft Copilot pricing
For individuals, Microsoft’s current power-user offer is Microsoft 365 Premium at $19.99 per month, combining Microsoft 365 benefits with Copilot capabilities. It covers one to six people for the Microsoft 365 apps, but the AI benefits are for the subscription owner. Microsoft lists Copilot in Word, Excel, PowerPoint, and Outlook, along with higher usage and agents for tasks such as source-cited research and data analysis.
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 plan, SuperGrok at $30 per month, and SuperGrok Plus at $100 per month. SuperGrok adds Grok 4.5, higher limits, image and video generation, connectors, and Expert. SuperGrok Plus adds significantly higher usage, 1080p video creation, faster replies, priority at peak times, and early feature access.
xAI’s comparison table also names Lite and Heavy tiers, but the public page did not expose stable US list prices for them during this check. We therefore exclude those prices instead of relying on checkout screenshots or old reports.
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, and verified students or educators can get Education Pro for $10 per month. Pro includes advanced models, extended research, file analysis, image and video generation, and limited file-and-app creation. Max raises those limits and adds early access and Perplexity Computer features such as Brain in research preview.
The important catch is the separate Computer credit system. Perplexity says consumer Pro does not include a recurring monthly Computer credit allocation after any one-time bonus, while consumer Max includes 10,000 credits per month. Computer work can therefore cost more than the subscription price. Search and ordinary Ask queries do not consume those credits.
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, Perplexity Computer credits
Mistral pricing
Mistral lists Pro at $14.99 per month, a verified higher-education student plan at $5.99 per month for eligible new users, and Team at $24.99 per user per month. Taxes are excluded. The student offer lasts a maximum of 12 months and has eligibility rules.
Mistral now positions the paid plan around Vibe as well as chat. Pro includes higher chat and search limits, all-day coding in the command line, development environments, and the web, plus more image generation. It deserves consideration when price, coding access, European provider context, or Mistral’s ecosystem matters more than having the largest consumer bundle. Source: Mistral pricing
Cursor pricing
Cursor offers a limited Hobby plan, Pro at $20 per month, Pro+ at $60, and Ultra at $200. Its current pricing page emphasizes Agent limits, frontier models, MCPs, skills, hooks, cloud agents, Composer, and usage-based Bugbot billing. Paid tiers provide 3x or 20x the Pro Agent limits, but actual value still depends on which models and workflows you use.
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 now lists $15 in total monthly AI credits for Pro, $70 for Pro+, and $200 for Max. Chat, agent mode, code review, cloud agents, the command-line tool, and Copilot Apps can consume those credits, so older “premium request” counts are 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 paid AI subscription is worth it?
There is no defensible universal winner. Use the decision below:
- Pick Google AI Plus when $7.99 is your ceiling and its Gemini, Gmail, Notebook, Flow, storage, and family-sharing bundle is available where you live.
- Pick GitHub Copilot Pro when you want the lowest-priced mainstream coding subscription inside familiar development tools.
- Pick Mistral Pro when you want a lower-priced general assistant with integrated coding access.
- Pick ChatGPT Plus for the broadest all-purpose workspace.
- Pick Claude Pro when writing, analysis, Claude Code, or its growing suite of specialized work tools 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 in the August 11 update
- Google AI Plus created a new $7.99 step between free and Pro, while Google’s wider bundle added more product benefits and regional caveats.
- ChatGPT Free, Plus, and Pro now differ across a much wider set of agent, coding, research, media, and productivity features.
- Claude Pro now bundles Code, Cowork, Design, Science, Research, and Microsoft 365 access.
- xAI publicly lists SuperGrok Plus at $100, but not stable prices for every named Grok tier.
- Perplexity’s Computer tasks now require separate credits; Pro has no recurring monthly allocation, while Max includes 10,000 credits.
- Mistral’s subscription value now includes its Vibe coding environment, not only the chat assistant.
- GitHub Copilot’s $10, $39, and $100 tiers now advertise monthly AI-credit values alongside features.
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. Consider the $7.99–$14.99 options before assuming $20 is the minimum. Pay for one product only after it earns a permanent place in your weekly workflow. Treat $100–$200 plans as capacity purchases that need a business case, and watch for separate agent or usage credits that can increase the real cost.
Always check the official provider page before checkout. The best pricing guide is a decision framework plus current sources—not a promise that fast-changing limits will stay fixed.