In short: As of July 2026, the cheapest paid entry point is Gemini’s Google AI Plus (₩7,500/month, roughly $5). ChatGPT starts at Go for ₩13,000 (about $9), and Claude’s Pro runs $18–22 a month. If you also want coding and work automation, Claude Pro — which bundles Claude Code — has the edge; for general chat and images, ChatGPT Plus; and for the lowest-cost paid entry and video generation, Gemini.
Try to pick one AI subscription and you’ll get a headache bouncing between three companies’ pricing pages. Every company names its tiers differently, some prices show in Korean won and others in dollars, and a few plans don’t even display a price on the page. So for this post I connected directly from Korea to the official ChatGPT, Claude, and Gemini pricing pages on July 7, 2026, captured the screens, and built this comparison from the displayed prices alone. These aren’t numbers copied from somewhere else — they’re the figures you see one screen before checkout.
Contents
- How I measured
- The three-way price table at a glance
- ChatGPT — the arrival of Go, and the ad clause
- Claude — the subscription includes Claude Code
- Gemini — the lowest entry price and the bundles
- Recommendations by use case
- My own subscription mix
- FAQ
1. How I measured
This tool comparison follows three rules. First, I use only the displayed prices confirmed by connecting directly, on July 7, 2026, to each company’s official pricing page (openai.com, claude.com, gemini.google, one.google.com). Second, I cover only personal consumer plans (Business and Enterprise excluded). Third, any number not shown on the page I mark as “not listed” rather than estimate. The USD figures in parentheses are rough conversions at approximately ₩1,400 to the dollar as of that date, offered only as a reference for readers outside Korea.
One thing I noticed while measuring, up front: even on the same day, the displayed price could differ depending on the connection environment. Claude in particular showed Pro as $17 in one browser environment and $18 in another. So read every number in this post as “what that screen displayed on that day,” and make your final check at your own checkout screen.
2. The three-way price table at a glance
| Tier | ChatGPT | Claude | Gemini (Google AI) |
|---|---|---|---|
| Free | Free ₩0 | Free $0 | Free ₩0 (15GB storage) |
| Entry | Go ₩13,000/mo (~$9) | — | AI Plus ₩7,500/mo (~$5) (400GB) |
| Mid | Plus ₩29,000/mo (~$21) | Pro $18/mo (annual billing, $220 upfront) / $22 if billed monthly | AI Pro ₩29,000/mo (~$21) (5TB) |
| Top | Pro from ₩159,000/mo (~$115) (5x or 20x limits) | Max from $110/mo (5x or 20x vs Pro) | AI Ultra ₩119,000/mo (~$85) (5x) / ₩300,000/mo (~$215) (20x, 20TB+) |
Even from the table alone, the structure is clear. The cheapest paid entry is Gemini AI Plus (₩7,500); the mid tiers converge for all three companies in the high-₩20,000s to around ₩30,000; and the top tiers spread from ₩110,000 all the way to ₩300,000. It’s also worth noting that all three top tiers are built on a “5x or 20x usage” structure.
3. ChatGPT — the arrival of Go, and the ad clause
What stands out in ChatGPT’s pricing is Go (₩13,000/month, ~$9), slotted in between Free and Plus. It’s a light paid tier that raises the GPT-5.5 Instant usage limit and lifts the caps on messages, uploads, and image generation — but the page spells out a clause that reads “this plan may include advertising.” The condition that a paid plan can still carry ads is something to know before you subscribe.
From Plus (₩29,000/month, ~$21) up, you unlock advanced reasoning based on GPT-5.5 Thinking, expanded limits on deep research and agent mode, and projects, scheduled tasks, and custom GPTs. Per the official comparison table, the reasoning-model context window is 256K on Plus and 400K on Pro (from ₩159,000/month, ~$115), and Pro adds GPT-5.5 Pro reasoning, unlimited image generation, and Codex up to its maximum limit. Pro’s “from” reflects the choice between 5x and 20x limits.
In short, ChatGPT has the smoothest lineup. You can start in the low-₩10,000s and step up in stages, but the peak performance of the coding agent (Codex) is concentrated in the top tiers.
4. Claude — the subscription includes Claude Code
Claude keeps it simple with three tiers: Free, Pro, and Max. Pro showed as $18/month on annual billing ($220 upfront), or $22 on monthly billing. As noted above, $17/$20 markings were also observed depending on the connection environment, so confirm at your checkout screen. Max starts at $110/month and offers 5x or 20x the usage of Pro (the exact price of the 20x tier is shown only as a starting figure on the page).
Claude’s real differentiator sits outside the price table. From Pro up, Claude Code (a terminal coding agent) and tools like Cowork, Design, and Science are included in the subscription. For someone who wants coding automation, that means you don’t have to budget for “a chatbot subscription plus a separate coding-tool subscription.” The model lineup is Fable, Opus, Sonnet, and Haiku, and access to the higher models widens with your tier. The recent change of Claude Code’s default model to Sonnet 5 is laid out in a separate post.
The thing to watch is the billing currency. Of the three companies, only Claude was displayed in dollars even on Korean access. The won-converted amount shifts with the exchange rate and VAT, so if you’re comparing on a fixed won budget you’ll need to account for that variability.
5. Gemini — the lowest entry price and the bundles
The strength of Gemini (Google AI) pricing is its entry price and its bundles. Google AI Plus is ₩7,500/month (~$5), the cheapest of the three companies’ paid plans, and it comes with 2x the free usage, video generation, 200 Flow credits, expanded NotebookLM, and 400GB of cloud storage (shared across Gmail, Drive, and Photos). For anyone who needs the storage in its own right on top of the AI features, the perceived value climbs even higher.
Google AI Pro (₩29,000/month, ~$21) provides 4x the free usage, 1,000 Flow credits, 5TB of storage, and 5x the NotebookLM limit. Note that this price isn’t shown on the gemini.google pricing page but is confirmed on the Google One pricing page. AI Ultra comes in two steps — ₩119,000/month (~$85; 5x vs Pro) and ₩300,000/month (~$215; 20x) — with early access to advanced features like Deep Think and 20TB+ of storage (some features are noted as limited to the US region and English).
The model wording is described, on the free tier, as “access to 3.5 Flash, and graduated access to 3.1 Pro.” If a large share of your AI use happens inside the Google ecosystem — Search, Gmail, Docs — it has more touchpoints than a same-priced subscription elsewhere.
6. Recommendations by use case
- Starting a paid plan for the first time: Gemini AI Plus (₩7,500). The lightest cost, with 400GB of storage thrown in.
- General chat, documents, and images: ChatGPT Plus (₩29,000). The widest spread of general-purpose features — reasoning models, deep research, custom GPTs. To trim the budget further, Go (₩13,000) works too if you can accept the ad clause.
- Coding and work automation: Claude Pro ($18–22/month). Claude Code is included, so there’s no separate coding-tool cost. Which model to pick is covered from a performance angle in the Opus 4.8 vs GPT-5.5 vs Gemini 3.5 comparison.
- Video generation and Google-ecosystem users: Gemini AI Pro (₩29,000) or above. Flow credits, NotebookLM, and the storage bundle create real practical value.
- Heavy users with high usage: All three companies’ top tiers (Max, Pro, Ultra) are “5x/20x” structures. If you frequently see limit-exceeded warnings, that’s the moment to move up, and it won’t be too late.
7. My own subscription mix
I pay for Claude Max and work in Claude Code every day. The publishing-automation pipeline behind this very blog runs on top of Claude Code, and work footage like the WordPress internal-link automation is its output too. Because the subscription call was a tough one, I first worked through a Max subscription vs API key cost comparison, and the conclusion was that at my daily usage a subscription is a predictable fixed cost.
On top of that I run two free plans in parallel. Search and light questions go to Gemini free, and ChatGPT free serves for checking specific features. With one paid plus two free, I can keep my monthly fixed cost concentrated in one place while still tracking the changes across all three ecosystems — which, for cost management relative to productivity, has been the smoothest setup so far.
8. FAQ
Q1. Can’t I just get by on free?
All three companies have free plans, and the quality isn’t low. But on all three, the free-tier caps on messages, images, and research are noticeably tight. Once “work got cut off because of a limit” repeats more than once a week, that’s your cue to upgrade.
Q2. Is it worth subscribing to two or more?
Only when the use cases clearly diverge. A non-overlapping combination — say, Claude for coding and Gemini for video generation — makes sense, but pay for two with the same purpose and one will mostly sit idle.
Q3. What if this post’s prices differ from my actual checkout price?
That’s normal. Displayed prices vary by region, exchange rate, promotion, and billing cycle (monthly/annual), and even during my own measurement I observed different displayed prices on the same day. The numbers in this post are as of Korean access on July 7, 2026; for the final price, confirming at each company’s checkout screen is the accurate move.
Prices can change at any time. Before subscribing, be sure to check the current price on each company’s official pricing page (sources below).
Sources (captured on the Korean pricing pages, 2026-07-07): OpenAI — ChatGPT pricing, Claude — pricing, Gemini — subscriptions, Google One — AI plans