Reference · Updated Quarterly

AI Tool Pricing Comparison 2026

Last updated: May 24, 2026 · Next update: August 24, 2026

What this page is

A side-by-side pricing comparison of the eight major consumer AI tools OneHuman tracks: ChatGPT, Claude, Gemini, Copilot, Perplexity, Grok, Codex, and Cursor. Updated quarterly from the OneHuman pricing sweep — not from press releases, not from vendor marketing copy, not from third-party screenshots. The number in each cell is the number on the vendor's pricing page on the date this page was last regenerated.

If you arrived here from a search engine asking "what does ChatGPT Pro cost vs Claude Max" or "compare AI tool pricing 2026," this page exists for that question.

The eight tools

The OneHuman watchdog tracks eight consumer-facing AI tools as of 2026. The list reflects where the consumer money is actually going, not where the venture capital is going. Inclusion does not imply endorsement; tools are tracked because consumers are paying for them and need pricing transparency.

Tool Vendor Category
ChatGPT OpenAI General-purpose assistant
Claude Anthropic General-purpose assistant
Gemini Google General-purpose assistant, Workspace-integrated
Copilot Microsoft General-purpose assistant, Office-integrated
Perplexity Perplexity AI Answer engine / research
Grok xAI General-purpose assistant, X-integrated
Codex OpenAI Code-specific assistant (mobile + cloud)
Cursor Anysphere Code editor + AI assistant

Tier 1: Free tiers compared

Every tool on the list has a free tier. The question is what "free" means — what you get, what's withheld, and what data trade you're making in exchange.

Tool Free tier cost Key inclusions Key restrictions Data trade
ChatGPT Free $0 Access to default model, web browsing, file uploads (limited) Reduced rate limits, no priority on new releases Conversations may be used for training unless explicitly opted out in settings
Claude Free $0 Access to Claude default model, web search (limited) Daily message cap, no Projects [VERIFY FROM SWEEP DATA]
Gemini Free $0 Neural Expressive UI, base Gemini model, basic image generation No Daily Brief, no Spark, no Omni video Standard Google data terms apply
Copilot Free $0 Web chat, basic image generation No GPT-5/equivalent model, no Office integration [VERIFY FROM SWEEP DATA]
Perplexity Free $0 3 Pro Searches per day, default model Most searches use standard model, limited file uploads [VERIFY FROM SWEEP DATA]
Grok Free $0 Limited Grok 3 access, on X with rate limits Heavy rate limits, no SuperGrok features X profile data integrated by default
Codex Free $0 Mobile app + cloud Codex, code generation Free tier code processed under default OpenAI data terms Code may be used for training unless opted out
Cursor Free $0 2,000 completions/month, limited premium model requests Slow request queue at peak times [VERIFY FROM SWEEP DATA]

The watchdog read on free tiers: Free is real but it is not unconditional. The data trade is the price. For personal use, casual queries, and learning, every tool's free tier is acceptable. For anything proprietary, regulated, or under NDA, free tier defaults are not the same as paid tier defaults — read the data-access comparison page (/reference/ai-tool-data-access-comparison-2026) before pasting work content into any free tier.

Tier 2: Mid-tier paid plans ($15–25/month range)

The competitive middle. Almost every tool prices its mid-tier around $20/month, which is no accident — it is the price ChatGPT Plus established in 2023 and the rest of the market has matched. Comparing what $20/month buys you across vendors is the most useful comparison for most consumers.

Tool Mid-tier name Price What you get vs free Best for
ChatGPT Plus Plus [VERIFY] ~$20/mo Higher rate limits, priority on new models, GPTs, DALL-E, voice Most consumers doing meaningful work in one tool
Claude Pro Pro [VERIFY] ~$20/mo 5× more usage, Projects, priority access Writers, analysts, long-context work
Gemini → Google AI Google AI [VERIFY] ~$19.99/mo Daily Brief (Gmail/Calendar digest), Gemini Omni video, expanded model access Google Workspace users
Copilot Pro Pro [VERIFY] ~$20/mo Priority access, advanced models, image generation Bing-search-heavy users
Perplexity Pro Pro [VERIFY] ~$20/mo Unlimited Pro Searches, choice of underlying model, file uploads Research, citation-heavy work
Grok Premium+ Premium+ [VERIFY] ~$16/mo Higher Grok limits, ad-free X X power users
Codex Bundled in ChatGPT Plus (see ChatGPT Plus) Higher code generation limits Developers
Cursor Pro Pro [VERIFY] ~$20/mo 500 fast premium requests, unlimited slow requests Working developers

The watchdog read on mid-tier: Mid-tier is the right price point for committed users of a single tool. The price parity across vendors is real — picking based on $1–2 monthly differences is not the right axis. Pick based on which tool's strengths match your work. If you only use AI a few times a week, the free tier is sufficient and the $20/month is wasted.

Tier 3: Premium and unlimited tiers ($100–250/month range)

The premium-tier market is where vendor strategies actually diverge. ChatGPT Pro at $200, Claude Max at $100, Google AI Ultra at $249.99 — these are not equivalent products at different prices. They are three different bets on what a premium AI subscriber wants.

Tool Premium tier name Price The premium pitch The asterisk
ChatGPT Pro [VERIFY] ~$200/mo Unlimited access to o-series models, Operator agent, Sora video Genuine unlimited has been the most consistent of the premium tiers
Claude Max [VERIFY] ~$100/mo 5–20× more Claude Code usage, priority access to new models, Projects expansion Significantly cheaper than ChatGPT Pro for similar value to most users
Gemini Google AI Ultra [VERIFY] ~$249.99/mo Gemini Spark (24/7 background agent), Gemini Omni video, max model access Most expensive premium tier; Spark requires standing Gmail access (see data access comparison page)
Cursor Business [VERIFY] ~$40/user/mo Unlimited fast requests, team controls, SSO Designed for organizations; not consumer-facing premium

The watchdog read on premium tiers: Premium tiers are for people who can name the specific feature they need that the mid-tier does not deliver. If the answer is "I just want better AI" without specificity, the premium tier will not feel like a $100–200 upgrade over the mid-tier. If the answer is "I need unlimited Codex usage" or "I need Spark agent access" or "I need Sora video generation," the premium tier becomes worth the math.

Team and Enterprise tiers (per-user pricing)

Team plans price per user per month and add admin controls, SSO, and (in some cases) different data-use defaults. The comparison gets useful at 3+ users — below that, individual paid tiers are usually cheaper.

Tool Team plan name Price per user Minimum users Notable inclusion
ChatGPT Team [VERIFY] ~$30/user/mo 2 Admin console, default data-training opt-out for the team
Claude Team [VERIFY] ~$25/user/mo 5 Projects, central billing
Gemini Google Workspace Business AI add-on Varies by Workspace tier Per Workspace seat Integrated into Workspace admin
Copilot Microsoft 365 Copilot [VERIFY] ~$30/user/mo Existing M365 license required Office 365 deep integration
Cursor Business [VERIFY] ~$40/user/mo None stated Org-wide settings, audit

The watchdog read on team plans: Team plans solve real problems (billing, admin, data-policy defaults) and are worth it for organizations. They are not worth it as a "discount for individuals" — the per-user math at the minimum user count is usually higher than individual paid tiers.

Hidden costs that don't show on the pricing page

The vendor pricing pages tell you the headline number. They do not always tell you:

  • API overage. If you build a workflow that calls the API beyond the included quota, you pay per-token at rates that compound silently. Claude Code, Cursor, and Codex are common cases where heavy use can exceed the included usage.
  • Geo pricing variation. ChatGPT Plus in India is priced lower than ChatGPT Plus in the US. Gemini's Google AI plans vary by country. The prices quoted in this comparison are US-default prices.
  • Subscription stacking. Several products require an underlying subscription before the AI tier becomes available. Microsoft 365 Copilot requires M365. Google AI Ultra requires a Google account but no additional subscription. Cursor Pro is standalone.
  • Annual vs monthly. Some vendors offer 15–20% discounts on annual billing. The headline monthly number is the worst-case rate.
  • Promotional pricing decay. Launch promotions (e.g., 3 months at $5) are not the steady-state price. If a tool's pricing seems unusually low, check whether you are on a promo that re-prices at month 4.

The OneHuman verdict — which tier fits which user

For the casual user (a few queries per week, mostly personal): free tier of whichever tool's interface you prefer. Spending money here is pre-emptive, not warranted.

For the committed single-tool user (daily use, meaningful work, one tool dominates): mid-tier of that tool. $20/month is the right price point. Switching costs between tools at this tier are low; experiment.

For the multi-tool user (different tools for different jobs — Claude for writing, Codex for coding, Perplexity for research): two or three mid-tier subscriptions at $20 each. This is cheaper and more capable than one premium tier.

For the premium-justified user (specific feature requirement that only premium delivers — unlimited Codex, Spark agent, Sora video): the premium tier that delivers that specific feature. Do not buy the premium tier of a tool whose mid-tier already meets your needs — the premium price gap rarely matches the value gap for most use cases.

For the organization with 5+ users: team plan of the dominant tool. The admin controls and data-policy defaults are worth the per-seat premium over individual subscriptions.

Data currency note

The prices in this comparison are pulled from the OneHuman pricing-sweep persistence layer on the date stamped at the top of this page. AI tool pricing changes mid-quarter sometimes. The sweep runs every 15 days; this page regenerates quarterly. If a tool you care about has shifted pricing since the last update, the freshest reference is the vendor's own pricing page — linked from each tool's OneHuman hub page (/tools/{slug}).

Independent. AI-assisted. Human-verified. No ads. No affiliates. No investors.