I’ve spent the better part of this year testing ChatGPT alongside Claude and Gemini for actual client work — content research, coding sanity-checks, and drafting — so this isn’t a copy-paste of a press release. It’s the numbers, checked against primary sources, plus what I’ve actually experienced logging in and using the product day to day.

1. How Many Users Does ChatGPT Have in 2026?

ChatGPT’s growth curve is still the reference point every other AI product gets measured against. OpenAI officially confirmed 900 million weekly active users in late February 2026, alongside its $110 billion funding round — more than double the 400 million it reported a year earlier.

By late July 2026, The Information reported — citing internal OpenAI figures, not an official announcement — that weekly actives were nearing 1 billion, about seven months behind the milestone OpenAI had internally targeted for the end of 2025. Separately, Sensor Tower data cited by Reuters showed the ChatGPT app crossed 1 billion monthly active users in June 2026, making it the fastest consumer app in history to reach that mark.

Two figures are worth separating out because people conflate them constantly: weekly active users (a rolling count of unique people who opened the product in the last seven days) and monthly active app users (Sensor Tower’s app-store-based estimate). They measure overlapping but different populations, so it’s normal to see “900 million weekly” and “1 billion monthly app users” reported in the same month without contradiction.

India overtook every country except the US in growth terms, crossing 100 million weekly active users in early 2026 — a detail I found genuinely surprising given how much English-language coverage centers on the US market. If you’re weighing ChatGPT against alternatives for a specific workflow, our deep dive on ChatGPT for coding covers where it actually holds up under pressure.

2. Daily Prompts and Monthly Traffic

ChatGPT users send an estimated 2.5 billion prompts every day globally, with roughly 330 million of those coming from the US alone. Similarweb logged 5.7 billion total visits to chatgpt.com in a single month (January 2026) — the fourth-highest monthly figure the site has ever recorded, and enough to keep ChatGPT inside the top 15 most-visited websites on the internet.

3. OpenAI’s Revenue, Valuation, and the IPO Filing

This is where 2026 diverges sharply from the story a year ago. OpenAI’s CFO Sarah Friar confirmed annualized revenue passed $20 billion in January 2026, and by February that run rate had climbed past $25 billion. Q1 2026 revenue alone came in at $5.7 billion — but the company also posted a non-GAAP operating margin of negative 122%, meaning it spent roughly $2.22 for every dollar it brought in.

On June 8, 2026, OpenAI confidentially filed paperwork with the SEC for an IPO, targeting a valuation north of $1 trillion. As of this writing, reporting suggests the company is leaning toward pushing the listing into 2027 rather than rushing a 2026 debut, with Sam Altman reportedly treating anything below $1 trillion as a non-starter.

Note: sources report slightly different revenue figures for 2024–2025 depending on whether they count run-rate at year-end versus full-year totals. The direction and scale of growth is consistent across every source I checked; the exact quarter-by-quarter numbers are not.

OpenAI’s post-money valuation reached $852 billion after a $122 billion funding round closed in March 2026. Microsoft holds the largest disclosed outside stake at roughly 27% (worth an estimated $228 billion at that valuation), followed by SoftBank at around 11.7% and Nvidia at roughly 3.5%. For comparison against the coding-tool side of this ecosystem, our breakdown of GitHub Copilot’s 2026 credit system shows how differently competing AI vendors are pricing usage right now.

4. ChatGPT’s Market Share vs. Gemini, Claude, and Perplexity

Here’s something I want to flag honestly rather than paper over: ChatGPT’s “market share” in 2026 depends heavily on who’s measuring it and what they’re counting. I pulled figures from four different analytics firms and they don’t agree with each other — not because any one of them is wrong, but because “web referral share,” “app usage share,” and “total web visits” are three different things.

What every methodology agrees on, though, is the trend line: ChatGPT is still comfortably in first place, but its lead is shrinking. Similarweb’s numbers show ChatGPT’s worldwide web-visit share fell from about 79% a year earlier to roughly 54–58% by mid-2026, while Gemini’s share nearly doubled and Claude posted the fastest year-over-year growth of any assistant in the set — up around 855% year over year on the web.

5. Where ChatGPT Is (and Isn’t) Available

ChatGPT is accessible in roughly 170+ countries. Estimates on the blocked/restricted side range from about 20 to 25 countries depending on how you count outright government bans versus OpenAI-side infrastructure restrictions. The core government bans are consistent across every source I checked:

• China — blocked as part of the Great Firewall; domestic alternatives like DeepSeek and Qwen fill the gap
• Russia — blocked by Roskomnadzor; OpenAI also stopped serving Russian traffic directly starting in 2024
• North Korea, Iran, Syria, and Cuba — blocked under broader information-control or sanctions regimes

Italy briefly banned ChatGPT in March 2023 over GDPR concerns — the first Western country to do so — but lifted the ban within weeks after OpenAI added age verification and clearer data controls. As of 2026, all 27 EU member states have access, under what’s generally regarded as the strictest regulatory scrutiny of any bloc.

6. ChatGPT Pricing Plans (2026)

OpenAI now runs six subscription tiers. The headline monthly prices haven’t moved much this year, but what each tier includes has changed several times — most notably the April 2026 introduction of a $100 Pro tier sitting between Plus and the original $200 Pro.

Enterprise pricing stays custom and quote-based. Worth noting: Free and Go plan content is used to train OpenAI’s models by default, with an opt-out available — Business and Enterprise plans exclude your data from training automatically.

7. Which GPT Model Powers ChatGPT Right Now?

As of August 2026, OpenAI is rolling out the GPT-5.6 family — internally split into Sol (flagship reasoning), Terra (used inside Codex and ChatGPT Work), and Luna (the new default for Free and Go). Rollout varies by account and region, so don’t be surprised if your model picker looks slightly different from a colleague’s on the same plan.

A practical note from testing: “getting GPT-5.6” doesn’t mean every single message runs the largest, most expensive version of the model. Each response can use a different reasoning-effort setting, and the highest tier (“ultra”) is gated to Sol specifically. If your replies feel inconsistent in depth from one message to the next, that’s usually why.

8. Who Owns and Funds OpenAI?

OpenAI is still privately held, so there’s no ticker to check. But the cap table that leaked in early 2026 gives a reasonably clear picture of who benefits most from ChatGPT’s growth:

Sam Altman himself reportedly holds no direct equity in the company — a detail that surprises most people the first time they hear it. Microsoft’s commercial relationship also loosened in April 2026: Azure exclusivity ended, and OpenAI’s revenue-share obligations to Microsoft are now capped through 2030.

9. What People Actually Use ChatGPT For

OpenAI’s own economic-research analysis of millions of anonymized conversations still shows “practical guidance and asking questions” as the single largest use-case bucket, followed by writing assistance and then coding. Enterprise adoption has kept accelerating: 92% of Fortune 500 companies now use OpenAI’s tools in some form, and workplace seats have grown roughly 9x year over year to more than 7 million.

• Marketing and content teams — drafting, research, and SEO-adjacent writing (see our guide on writing SEO articles with AI)
• Software development — inline coding help, debugging, and increasingly, Codex-driven agentic tasks
• Customer support — ticket triage and drafted responses layered on top of existing helpdesk tools
• Education — tutoring, summarization, and study assistance, despite ongoing academic-integrity debates
• Finance and healthcare — internal research and drafting, generally kept away from regulated decision-making

10. My Experience Testing ChatGPT This Year

How I Test the Platforms I Review

I don’t review AI tools off a spec sheet. Every platform I cover on Websites2Know gets a real account, created the same way any reader would sign up — free tier first, then upgraded if the paid tier changes the story. I run the tool through actual workflows (research, drafting, coding sanity-checks, and image/data tasks where relevant), take notes on what breaks or surprises me, and only then write the piece you’re reading. This review reflects my personal experience and opinion. It isn’t professional, financial, or legal advice — for anything account-specific, billing-related, or compliance-related, go straight to OpenAI’s own documentation or support team.

Signing up in mid-2026 is noticeably faster than it was even a year ago. Google SSO gets you into a working chat window in under 15 seconds — no phone verification wall on the free tier unless you trip a usage-abuse flag, which I did once by testing rapid-fire prompts back to back, and got a one-time SMS check as a result. Minor friction, but worth knowing if you’re testing limits deliberately.

Here’s my honest take, and I’ll be blunt about it: ChatGPT is still the most capable general-purpose assistant I’ve used for breadth — it’s genuinely strong across writing, quick coding fixes, image generation through Sora integration, and voice mode. But it is not the tool I reach for first anymore for long, structured writing work. Claude’s outputs need less editing when I’m producing anything over 1,500 words, and I say that as someone who spends most of a working day inside ChatGPT.

The rough edges I’ve hit personally: Deep Research on the Plus tier is capped at 10 runs a month, and I’ve burned through that allotment inside a single busy week more than once. The model picker can be confusing — GPT-5.6 Sol doesn’t run every message even when it’s technically “available” on your plan, and figuring out which reasoning-effort tier you’re actually getting takes some digging through settings most casual users will never find. I’ve also seen the same factual claim answered two different ways in the same session after asking a follow-up question — not constant, but frequent enough that I always verify anything with real stakes attached.

Agent Mode, which lets ChatGPT act semi-autonomously across a browser session, impressed me more than I expected going in — it handled a multi-step research-and-summarize task cleanly on the first attempt. But it’s slow relative to just doing the task myself for anything under about ten minutes of manual work, so I only reach for it on genuinely tedious multi-tab jobs now.

11. ChatGPT vs. Claude vs. Gemini vs. Perplexity: Quick Comparison

If you’re deciding between ChatGPT and Claude specifically for coding or writing workflows, I’ve written a more detailed side-by-side in our Claude Projects vs. ChatGPT GPTs comparison, and a dedicated look at whether Claude AI is safe and trustworthy.

FAQ

How many people use ChatGPT in 2026?

OpenAI confirmed 900 million weekly active users in February/March 2026, with unofficial reporting from late July putting that figure close to 1 billion. Separately, the ChatGPT app crossed 1 billion monthly active users in June 2026.

How much revenue does OpenAI make from ChatGPT?

OpenAI’s annualized revenue run rate reached roughly $25 billion by February 2026, though the company remains deeply unprofitable, with a Q1 2026 non-GAAP operating margin of negative 122%.

Is ChatGPT still the market leader among AI chatbots?

Yes, by every measurement source, though its lead has narrowed substantially in 2026. Depending on the methodology, ChatGPT’s share ranges from roughly 46% (app usage) to 77% (web referrals), with Gemini and Claude both gaining ground quickly.

What does ChatGPT cost in 2026?

Free ($0), Go ($8/mo), Plus ($20/mo), Pro ($100 or $200/mo), and Business ($20–25/seat/mo). Enterprise pricing is custom.

Is OpenAI going public?

OpenAI confidentially filed for an IPO on June 8, 2026, targeting a valuation above $1 trillion, but as of mid-2026 the company appears to be leaning toward delaying the listing into 2027 rather than 2026.

Which countries can’t access ChatGPT?

China, Russia, North Korea, Iran, Syria, and Cuba maintain outright government bans. Total blocked or restricted countries sit somewhere between 20 and 25, depending on the source and how “restricted” is defined.

Conclusion

ChatGPT is still the biggest name in AI by almost any measure that matters — users, revenue, and raw brand recognition. But 2026 is the year the numbers stopped being a one-horse race. Weekly actives are nearing a billion, revenue has more than doubled since last year, and an IPO north of $1 trillion is realistically on the table. At the same time, Gemini and Claude are eating into ChatGPT’s share faster than at any point since launch, and the day-to-day experience — at least in my testing — shows real cracks in consistency and Deep Research limits that heavier users will hit fast. If you’re choosing a primary AI assistant in 2026, ChatGPT remains the safest default for breadth and multimodal work. It is no longer the automatic best choice for every single workflow, and treating it that way in 2026 would be a mistake.