What IP Purity Score Counts as Clean? How to Read Each Tier
To judge what IP purity score counts as clean, just look at the risk score from your test: under 20 is Pure, the safest tier for registering high-risk accounts; 20 to 34 is Fairly Pure, good enough for everyday account care and routine operations; once you hit 50 or above, you're in Suspicious or High Risk territory, and for sensitive work you should switch lines outright. The lower the score, the cleaner the exit — but when you're picking a line, you still need to look at the sharing count and blacklist record alongside it.
When you get a new line, most people's first instinct is to open ISPing.me and check it once on the homepage. A number pops up on screen, and then the doubt sets in: does this score actually count as clean, and will it get blocked if I use it to register an account? Once you understand the conversion rules and the tier standards behind the number, you'll never be left staring at it wondering again.
Purity = 100 - risk score
People new to testing tools often get the concept backwards and assume a higher score means better quality. In network security and anti-fraud, the industry scoring logic works the other way around: the system counts how many negative labels an address is carrying — in other words, its risk score.
You can understand the relationship with one simple formula: purity = 100 - risk score. Say your test result shows a risk score of 12 — that means it has barely left any record behind, which converts to a cleanliness level of 88. Flip it around: if the risk score shoots up to 78, its cleanliness is only 22, a heavily polluted state.
The ISPing.me homepage shows the score from the risk side directly, on a scale from 0 to 100. The closer the number is to 0, the more comfortable the target site's firewall feels about that exit; the closer it creeps to 100, the more abnormal records it has racked up across the major threat-intelligence databases. Once you're clear on the direction, you won't mistake high risk for high purity when you're reading a report. For a breakdown of the underlying scoring weights and the historical items that cost points, see What Is IP Purity? How to Read the Risk Score.
The five-tier table
To make it easy to check at a glance, ISPing.me splits the 0-to-100 scale into five clear tiers. You can take whatever test result you have and match it straight to a tier:
| Score Range | Purity Tier | Risk Characteristics | How Usable It Is for Real Work |
|---|---|---|---|
<20 | Pure | Almost no abuse history; network behaviour is close to a real home user | Good for registering new accounts, linking cross-border payouts, and highly sensitive work |
<35 | Fairly Pure | Occasional minor fluctuation, but the underlying credit is still healthy | Good for daily social media work, maintaining old accounts, and routine overseas office work |
<50 | Fair | Some history of past use, or signs of light sharing by multiple people | Sensitive platforms occasionally throw a captcha; best paired with a separate fingerprint environment |
<70 | Suspicious | Clear history of abuse, or exposed commercial-server attributes | Registering a new account is very likely to be rejected; frequent activity on an old account easily triggers a security lock |
≥70 | High Risk | Deep contamination, heavy concurrent crowding, or already on a public blacklist | The vast majority of commercial sites block it outright, and continuing to use it very easily gets the account banned |
The feel of each tier in practice is very different. Lines in the <20 Pure tier usually come from well-managed residential broadband pools, and opening overseas sites of all kinds is mostly smooth sailing. In the 20 to 34 range — the <35 Fairly Pure tier — the previous user most likely just did normal browsing or light logins and never triggered a platform ban, so it still runs very smoothly.
Once the score crosses 35 into the <50 Fair tier, the risk system's guard goes up noticeably. Browsing with an exit in this tier is fine, but if you fill in a form to register a new account, the page will often throw up a tricky image captcha or even ask for a second phone verification. As for the 50 to 69 range — the <70 Suspicious tier — and 70 and above, the ≥70 High Risk tier, it's basically a label on your forehead saying "proxy traffic", and any even slightly strict site will cut the request off at the connection stage.
What score do you need for each use?
Talking about scores without a concrete use case is meaningless — streaming videos and linking a bank card demand completely different things from an exit. Match the pass mark below to whatever you're actually doing.
AI tool registration and overseas payment work have the strictest bar. When you're registering for ChatGPT or Claude, or linking Stripe or PayPal payouts, the IP purity of your exit should be kept strictly in the <20 Pure tier. Plenty of users try to save money by linking a card on a <50 Fair-tier node, only to have the system flag the transaction as high risk and freeze the account the moment they type in the card number. Payment risk engines are extremely sensitive to abnormal exits — even a dozen extra risk points can multiply the rejection rate, as explained in PayPal & Cross-Border Payment IP Risk Controls: Why You Keep Getting Flagged.
Running overseas social media and short-video platforms, you can relax the bar by half a tier. Maintaining a US TikTok account or running an Instagram matrix, go for the <20 Pure tier first, and if the budget is tight, the <35 Fairly Pure tier will still run stably. As long as you don't jump regions frequently, video reach and direct messages generally won't be restricted. But if you drop into the <50 Fair tier, newly posted videos will very likely get stuck at 0 plays, with the system lumping the account into a marketing-account observation pool — see What Does TikTok Require From an IP? How IPs Relate to Throttling and Bans.
If you're just reading overseas technical material day to day, watching streaming video, or running public web scraping that doesn't need a logged-in session, even a <50 Fair-tier line — or part of the <70 Suspicious tier — can just about cope. These scenarios don't involve any account assets building up, so at worst you'll click through a few extra human verifications while searching. As long as you don't hit the ≥70 High Risk kind of outright port blocking, it barely affects everyday use. What streaming unblocking requires from an IP, see Why Can't I Watch Netflix? How IPs Affect Streaming Unblocking.
A low score doesn't mean absolutely safe
Some people test an IP, get a score of just 15 — comfortably in the <20 Pure tier — happily use it to register a new account, and get banned in seconds anyway. The mismatch happens because a single score only reflects historical statistics, so when you're checking things over, you have to cross-check three other hard indicators alongside it.
The first thing to watch is the sharing count. The test page splits concurrent users into five tiers: 1-10 is Excellent, 10-100 is Good, 100-500 is Poor, 500-5000 is Very Poor, and reaching 5000+ is the Dangerous red line. Some providers have just launched a batch of new ranges, and the threat-intelligence databases haven't had time to log their abuse yet, so the IP risk score at that moment looks like it's only in the teens. In reality, more than eight hundred people are using that same exit to hammer APIs at the same time, and the moment you connect, the target site will block you for the concurrent density alone. How many sharing users is too many and when should you switch? See How Many Sharing Users Is Too Many? When Should You Switch IPs?.
Next, check whether the IP has been blacklisted. Different databases sync on different schedules, so an address might have been blacklisted by an anti-spam organisation yesterday for sending spam, while the overall score hasn't yet caught up and pushed the weight up. As soon as the blacklist status lights up red, mail providers and e-commerce risk systems will reject your requests outright. Can a blacklisted IP be cleaned, and how long does it take? See Can an IP on a Blacklist Be Cleaned? How Long Does It Take?.
Finally, check whether the network type is a native residential IP or a datacenter IP. Some ranges from lesser-known cloud providers don't get abused much in normal times, so their test scores can also sit in the <20 Pure band. But as long as the ASN is registered under a cloud computing company, the firewalls at banks and e-commerce platforms can tell at a glance that it's a datacenter IP and will still trigger strict blocking. So when you're picking an exit, only a line that combines a low risk score, a 1-10 Excellent or 10-100 Good sharing count, a clean blacklist record, and residential broadband attributes counts as genuinely clean. How to check ASN ownership and which numbers to watch out for, see What's the Deal With a Proxy IP's ASN?.
FAQ
Is a purity score of 25 clean? Can I use it to register an account?
A score of 25 falls in the <35 Fairly Pure tier, with an overall good reputation. It's plenty for logging in to maintain an old account or browsing overseas social media. But if you're registering a new account on a platform that's extremely sensitive to environment, or linking a payment card, go for a <20 Pure-tier residential line first — the pass rate will be much more reliable.
It was 15 (Pure) yesterday — why did it suddenly become 65 (Suspicious) today?
That usually means you're on a shared pool, and within the last 24 hours another user fired off bulk registrations or malicious crawling from the same exit, triggering a real-time alert in the threat-intelligence databases and pushing the risk score up fast. When that happens, switch to a dedicated node.
Does an IP risk score of 0 mean a site will never flag me?
No. A low risk score only means the exit itself has no dirty history. If your local machine has a DNS leak, its system timezone doesn't match the exit's city, or several accounts share the same browser fingerprint, the target platform will still judge the environment abnormal and trigger its risk controls.
A datacenter IP tested under 20 — can I use it as a residential IP?
No. Even if a datacenter IP has no abuse record for now and its score is low, its underlying ASN still belongs to a commercial datacenter. Banks, payment gateways, and mainstream e-commerce platforms recognise the server attributes directly and block it — it can't stand in for real residential broadband.
Notes
- Don't decide on the total score alone: after finding a
<20Pure score, always check that the sharing count is in the1-10or10-100range and that the blacklist is clear — only put it into real work once all three pass. - Test a new line with a throwaway account first: different platforms each keep their own private blocklists, so even with a clean residential IP, run a registration or browsing flow with a test account before logging in to your core main account.
- Keep your exit fixed — don't hop around: once you've picked a clean line, stick with it for the long term. Switching back and forth between several low-scoring nodes in a short time can send the account into a risk review for abnormal location changes.
Check your IP purity
Enter an IP on ISPing.me to see its risk score, purity, sharing count, and all other risk signals in detail.
Check nowRelated reading:What Is IP Purity? How to Read the Risk ScoreWhat Is the IP Sharing Count, and Why Does It Matter?What to Do If Your IP Is Blacklisted: How to Check and Remove ItResidential IP vs. Datacenter IP: What's the DifferencePayPal & Cross-Border Payment IP Risk Controls: Why You Keep Getting FlaggedWhat Does ChatGPT/Midjourney Require of Your IP?What Does TikTok Require From an IP? How IPs Relate to Throttling and Bans