Total — unique visitors
Browse by category Chatbots Image Generation Video Generation Audio & Voice Coding Writing Productivity Research AI Agents Free Tier Table
Home page ›Ask Cat on AI

Ask Cat › AI Tool Summary › ChatGPT

[Review] ChatGPT Pricing: Taiwan's NT$270 Go Plan Comes With an Ads Note the Official Page Buries

🐾 Key numbers
  • Free tier: There is
  • Cheapest paid plan: US$8/mo and up
  • Free quota: Free plan NT$0 per month for everyone (read directly from …
  • Last checked: 2026-09-08

Article last updated: 2026-10-07

ChatGPT’s own pricing table lists the Go plan at US$8 a month. Open the Taiwan version of the same page and Go actually bills at NT$270 a month — a real listed price we read directly off the site, not a currency conversion we calculated ourselves. Almost everything else about Go reads like a straightforward volume upgrade over free. One detail doesn’t: the plan’s own footnote says it “may include ads.” That line sits in the fine print, not in the feature highlights. If you’re comparing ChatGPT pricing before picking a plan, that’s the detail worth reading twice.

What “free” actually includes

Free costs NT$0 a month and anyone can sign up — we checked this directly on the Taiwan site on 2026-09-25. The free tier gives unlimited text chat on GPT-5.6 Luna (rolled out since 2026-08, still subject to anti-abuse limits), plus capped amounts of messaging with tools, file uploads, voice chat, image generation (described as slower on free), Deep Research, memory, Codex, and the desktop ChatGPT Work app.

The official plan-comparison table gives a few hard numbers. A “context window” is how much conversation the model can hold in its head at once, measured in tokens (small chunks of text). Free tier’s instant model gets a 27K-token window, versus 54K on Go/Plus and 128K on Pro. Free tier’s single-message input cap is about 12 pages of text, versus roughly 40 pages on Go/Plus and roughly 250 pages on Pro. For the reasoning model, free tier’s context window is listed only as “depends on circumstances,” while Go/Plus get 256K and Pro gets 400K. Response speed on free is described as “limited by system resources and service status”; paid tiers are simply labeled “fast.”

Independent testing behind the free tier (separate from the official comparison page) adds detail the vendor page doesn’t give: on GPT-5.3, free users get roughly 10 messages per rolling 5-hour window before being auto-downgraded to GPT-4o mini, which has no clear cap. Other free caps found in testing: about 2-3 generated images per day (varies with server load), 3 file uploads per day, and 16K of chat memory. The quota doesn’t reset at midnight — it’s rolling, so a message only frees up its slot after its own 5-hour window passes.

AMPM chart: ChatGPT (numbers taken from AMPM verified data, last verified 2026-09-08)

Figure: compiled by AMPM from its verified data (last verified 2026-09-08).

The Go plan: more quota, and a line about ads

Go runs GPT-5.2 Instant and, according to the vendor’s own plan breakdown, gives roughly 10 times the free tier’s chat and upload quota, plus the longer 54K context and ~40-page input limit already mentioned. It also raises the image-generation quota. What Go does not include: the advanced reasoning model (GPT-5.5 Thinking), task management, custom GPTs, or Deep Research — those stay locked to Plus and above. In plain terms, Go buys more room on the same basic model; it doesn’t buy a smarter one.

And then there’s the ads line. Go’s own feature notes state the plan “may include ads.” It isn’t framed as a tradeoff anywhere near the headline price — it reads as an afterthought next to a list of quota increases. For a plan priced at NT$270 a month in Taiwan, that’s a real condition to weigh, not a footnote to skim past.

Plus, Pro, and names that don’t match the live page

Plus costs US$20 a month (NT$690 in Taiwan, verified 2026-08-09) and runs GPT-5.5 Thinking — the reasoning model Go doesn’t get. Plus adds task management, custom GPTs, Deep Research, full productivity tools, and Sora video generation, with quota described as high enough that normal use rarely hits a ceiling.

AMPM illustration: The entry paid plan dramatically expands paper capacity, yet relies on a basic model without the deep

Illustration: The entry paid plan dramatically expands paper capacity, yet relies on a basic model without the deeper thinking brain.

Pro is listed at US$100 a month, with a Taiwan price of NT$3,300 (verified 2026-08-09), and runs GPT-5.5 Pro with roughly 20 times Plus’s usage allowance. Here the record gets messy: a third-party source claims a lower-priced Pro tier was added in April 2026, but we could not confirm that directly on the official page. A direct scrape of the official site on 2026-07-26 showed a single listing — “Pro: From US$100/month, 5x or 20x more usage” — with no “Pro Max” branding anywhere. A separate US$200 “Pro Max” entry in our data may simply be that same Pro plan’s 20x-usage option, not a distinct product. We’re flagging this as unresolved rather than guessing.

Business is listed at US$25 per seat (one paid user license) per month, also unconfirmed directly on the main pricing page. A secondary source — a summary on help.openai.com — suggests the real structure might be US$20 per user monthly on an annual plan, or US$25 per user monthly billed month-to-month, with a two-seat minimum. Treat both Pro’s branding and Business’s price as needing a direct check with OpenAI before you commit a company card.

Two time-limited offers are worth noting: ChatGPT for Teachers is free for verified US K-12 educators through June 2027, and nonprofits can get up to 75% off Business through OpenAI for Nonprofits. A separate promo that doubled Codex usage on the NT$3,300 tier ran only through May 31, 2026 — that window has already closed as of today.

🔍 AMPM exclusive check

The official comparison table never uses the word “ads” as a headline feature for Go — it’s buried in a secondary note, not the pricing grid. Our check confirms the line exists on the live page, just not where a skimming reader would look.

The vendor page also doesn’t explain what happens when a free user’s quota runs out mid-conversation. Independent testing found the model quietly drops down to GPT-4o mini without any clear on-screen warning — the conversation keeps going, just with a visibly weaker model. That’s a behavior the pricing page doesn’t document at all, since it only states quotas exist, not what happens after they’re hit.

One more gap: turning off personalized ads in account settings doesn’t just affect ads — it further shrinks the free quota to roughly 5 messages per 3-hour window, tighter than the standard ~10 per 5 hours. That’s a privacy/quota tradeoff the pricing page never mentions.

Finally, on the Pro/Pro Max naming question: our own proxy fetch of the official page on 2026-07-26 showed no “Pro Max” product at all, only one Pro listing with a 5x/20x usage choice. That’s a direct discrepancy between what some secondary listings claim and what the live page showed us.

☀️ AMO, the budget-minded cat

AMO, the budget-minded cat

Free is NT$0 and, on paper, generous: unlimited basic text chat is the headline. But testing shows the real ceiling is about 10 messages per 5 hours on the actual GPT-5.3 model before you’re bumped to GPT-4o mini, plus daily caps of roughly 3 file uploads and 2-3 images, and a 16K memory limit. For someone who chats with ChatGPT a few times a day, that’s plenty. For anyone running multiple tasks back-to-back, the 5-hour rolling reset bites fast.

Go’s NT$270 a month buys roughly 10 times that quota and a bigger context window (54K versus 27K), but no smarter model and no new features — it’s purely a volume upgrade, and it comes with that possible-ads condition. My take: if you’re not routinely hitting the free 10-message wall, don’t pay for Go just for headroom you won’t use. If you are hitting it constantly and don’t care about reasoning-model features, Go is a cheap fix — just go in knowing ads might show up. And if you’ve turned off personalized ads for privacy, know that it costs you quota too.

🌙 PIMI, the performance-minded cat

PIMI, the performance-minded cat

If you actually need ChatGPT to think harder, not just answer faster, Go won’t help — it runs the same instant model as free, just with more room. The real functional jump is Plus: it’s the first tier with GPT-5.5 Thinking, the reasoning model, plus a much bigger context window (256K on the reasoning model versus free’s “depends on circumstances”), Deep Research, custom GPTs, and task management. That’s a capability difference, not a quantity one.

Pro, at US$100 a month (NT$3,300 in Taiwan), adds GPT-5.5 Pro and about 20 times Plus’s usage allowance with a 400K context window — but the price gap is roughly 10x over Plus for mostly more headroom on an already-capable model. Worth it only if you’re hitting Plus’s ceiling daily; otherwise it’s paying for capacity you won’t use. One caution: the Pro/Pro Max split in some third-party listings doesn’t match what we found live on the official page, so confirm current terms before paying US$100 or US$200 for a specific tier.

Verdict: who should use it, who should skip it

Stick with free if you chat occasionally and can live with an unannounced downgrade to a weaker model once you hit roughly 10 messages in 5 hours. Go makes sense only if you hit that free ceiling often, don’t need the reasoning model, and are fine with a plan that may show ads for NT$270 a month — otherwise it’s an unnecessary step between free and Plus. Plus is the tier that actually changes what ChatGPT can do (reasoning model, Deep Research, custom GPTs) for US$20 a month (NT$690 in Taiwan), and is the one most people who outgrow free should land on. Pro is only for people already maxing out Plus daily — the 10x price jump buys mostly more usage, not more smarts. Treat Business pricing and the Pro/Pro Max naming as unconfirmed on the official page as of our last direct check, and verify both with OpenAI before paying for either.

Written: 2026-10-07 Price last verified: 2026-09-08 Tool last health-checked: 2026-10-07

📎 Sources

All prices and quotas in this article come from AMPM’s verified dataset for ChatGPT (last verified 2026-09-08); primary sources:

🐾 Meet the cats: AMO and PIMI

AMO

AMPM is a Taiwan-based AI-tool pricing watchdog. Our motto: ask before you subscribe. Two cats argue every tool from two sides:

  • ☀️ AMO — the budget-minded cat. Always asks: is the free tier enough, and is paying actually worth it?
  • 🌙 PIMI — the performance-minded cat. Cares about whether it works well, fast, and gets your job done.
PIMI

When they are done arguing, you know whether to pay. Every number is checked by us, with the verification date at the end.

🐾 Our sister sites also checked

Let's take a look at these

More verified articles on this tool

Go to the official website

Affiliate Links Notice