Direct answer
There is no single "good" Turnitin score, because Turnitin reports two different numbers with two different meanings. A similarity score is a percentage of your text that matched sources in Turnitin's database, and many universities treat roughly 10–24% as unremarkable for coursework, provided no single source dominates the report [1][2]. An AI writing score is only shown as an exact percentage when the detector's confidence clears its threshold; below that threshold Turnitin displays *%, which is a low-confidence signal rather than a clean result [3]. The reliable way to judge your own draft is to read the match overview and the AI indicator together, not to chase a target number [1].
What similarity percentage is considered a good Turnitin score?
Turnitin's own documentation is explicit that it does not define a good or bad similarity percentage, because the number measures overlap with a database, not the quality or honesty of your writing [1]. A 30% similarity can be perfectly legitimate if most of the matches are quotations, reference entries, or standard terminology from your field. Conversely, a 6% similarity can still contain a copied passage that an instructor would treat as misconduct, and the report would show that match clearly once you open it [1].
What matters far more than the headline figure is the distribution of matches. A report where one external source accounts for the bulk of the highlighted text is the pattern instructors scrutinise, because it suggests a single passage was lifted rather than incidental overlap with common phrasing [2]. A report where small percentages are spread across many sources — your bibliography, lecture slides, a textbook, a few journal articles — is usually just the normal fingerprint of academic writing [2].
Practical read: for coursework, many institutions regard roughly 10–24% as within normal range, but that is a convention, not a Turnitin rule, and thresholds vary by department and assignment type [2]. If your similarity sits above that band, the useful next step is to open the Match Overview and check whether the matches are properly quoted and cited. Fixing a missing citation is a five-minute edit; guessing whether the number is "too high" is not.
What does a good Turnitin AI writing score look like, and why does it show *%?
The AI writing indicator works differently from similarity. Turnitin only displays an exact percentage of text flagged as likely AI-generated when its detection confidence reaches its internal threshold; below that, the report shows *% instead of a number [3]. That asterisk is frequently misread. It is not a pass, a zero, or a guarantee that your writing is human — it means the detector did not gather enough signal to commit to a figure [3].
Because of this, a "good" AI score is less about a target percentage and more about confidence and context. A low exact percentage with the flagged segments concentrated in genuinely AI-assisted paragraphs is a far more coherent result than a *% that leaves you unable to see what was detected [3]. Turnitin also states plainly that the AI indicator is not proof of misconduct; it is one signal that instructors weigh alongside drafting history, version files, and their knowledge of your writing [3].
If your report shows a high AI percentage, the practical response is to look at which passages were flagged rather than the total. Heavily templated sections — generic introductions, restated definitions, transition-heavy conclusions — are the usual culprits, and rewriting them in your own voice with your own examples is both faster and more defensible than trying to game a detector [3].
How can I check my own Turnitin score before I submit?
Turnitin's student-facing guidance is that pre-submission checking is arranged by your instructor or institution; students normally cannot enrol themselves into the institutional database to run a draft [4]. Where a draft check is permitted, it is typically configured as a non-repository submission, meaning your draft is not added to Turnitin's student paper database and does not inflate your own similarity score later [4].
The value of seeing the report early is that it converts an anxious guess into a concrete edit list. You can see exactly which sources matched, whether a quotation is missing its citation, and which paragraphs the AI indicator flagged — and then fix those specific issues before the graded submission [4]. Turnitin's own advice to students is to use the report as a revision tool, not as a verdict [4].
This is also where an independent pre-submission check fits. A service like turnitin0 lets you upload your .docx, .pdf, or .txt and receive the same two PDFs your instructor sees — a Turnitin AI detection report and a similarity report — in under 15 minutes in 98% of cases. The check is non-repository, so your file is not added to Turnitin's student paper database, and English documents between 300 and 30,000 words are supported. That gives you the match overview and AI indicator described above on your own draft, while there is still time to act on them.
Reading your report is the point at which most students stop guessing and start fixing — and turnitin0 exists to put that report in your hands before your deadline does.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is 20% similarity bad on Turnitin?
Not by itself. Similarity counts matched text, not misconduct, and a 20% figure is within the range many universities treat as normal for coursework [1][2]. Open the Match Overview first — if the matches are quotations and references that are properly cited, the number is doing its job.
Why does my Turnitin AI score show *% instead of a number?
Turnitin only prints an exact AI percentage when its confidence reaches its threshold; below that it shows *% [3]. That asterisk means low confidence, not zero AI content, so treat it as "not enough signal" rather than a clean result.
Can a high similarity score get me accused of plagiarism?
A score alone is not an accusation. Instructors read the report — the source breakdown and the highlighted passages — alongside the assignment and your drafting history [1][2]. A single source accounting for most of the match is the pattern that draws scrutiny.
Can I check my Turnitin score before submitting?
Usually only if your instructor enables a draft check, since students cannot self-enrol in the institutional database [4]. Independent pre-submission services such as turnitin0 deliver the same AI and similarity PDFs, non-repository, so you can review the match overview and AI indicator before your real deadline.
Does a low AI score mean my writing is definitely human?
No. Turnitin states the AI indicator is one signal, not proof, and a low or *% result does not certify authorship [3]. Your drafting process, notes, and version history are what actually demonstrate how the work was produced.