By Sarina Gomez
This article is general information, not medical advice. Ask your midwife, OB-GYN or sonographer about your own pregnancy and scans.
No baby gender predictor is 100 percent accurate. The closest are medical: CVS and amniocentesis read the baby’s chromosomes and give an answer “with as much certainty as possible,” in ACOG’s words, but they’re done for medical reasons only. A cell-free DNA blood test from 10 weeks was right about 95 to 99 percent of the time in large reviews, and the anatomy scan is very accurate when the baby is in a good position. Urine kits, heart rate, the Chinese chart and other home tests perform at or near chance, so treat them as games.
Every test, rated
Here’s each method with the best published evidence we could find. “Correct” figures come from studies, not from the people selling a test.
| Test | When | Published accuracy | Verdict |
|---|---|---|---|
| Amniocentesis | 15–20 weeks | Reads the chromosomes directly | Medical use only |
| CVS | 10–13 weeks | Reads the chromosomes directly | Medical use only |
| Cell-free DNA blood test (in a clinic) | From 10 weeks | 95.4% for boys, 98.6% for girls (57 studies) | The most accurate non-invasive option |
| Anatomy ultrasound | 18–22 weeks | 99–100% after 13 weeks when the genitals are typical and visible | Very good, if the baby cooperates |
| Early ultrasound “nub” read by a sonographer | 12–14 weeks | About 92% at 12 weeks, 98–100% from 13 weeks in the best studies | Good in trained hands, not on a printout at home |
| At-home finger-prick DNA kit | Varies by kit | No independent peer-reviewed validation found; lab studies show blood tests before 7 weeks are unreliable | A guess with a lab attached |
| Urine “gender test” kits | Varies | Urine-based testing was unreliable in a JAMA review | Skip |
| Heart rate over 140 | Any | No meaningful difference between boys and girls | For fun only |
| Chinese gender chart | Any | No better than chance in 2.8 million births | For fun only |
| Ramzi theory, skull theory | 6 weeks, 12 weeks | Never confirmed in a peer-reviewed study | For fun only |
| Ring test, baking soda, cravings, bump shape | Any | Untested, or no link found | For fun only |
| Drano test | Any | No basis, and the chemicals are dangerous | Don’t do it |
The sections below explain each one and link the studies.
The medical tests: as close to certain as it gets
CVS and amniocentesis examine cells from the placenta or the amniotic fluid. ACOG describes them as tests that can tell you “with as much certainty as possible” about the chromosomes, done at 10 to 13 weeks (CVS) or usually 15 to 20 weeks (amnio). Both carry a small chance of pregnancy loss, so nobody has one to learn boy or girl. If you’re having one for medical reasons, you can ask whether the report includes the sex.
Cell-free DNA (cfDNA, often called NIPT) is a blood test from 10 weeks that screens for chromosome conditions and looks at the X and Y chromosomes (ACOG). A 2011 JAMA meta-analysis of 57 studies found blood-based testing detected boys 95.4% of the time and girls 98.6% of the time overall (Devaney et al.). A separate 2012 review of 90 studies and nearly 10,000 pregnancies put sensitivity at 96.6% and specificity at 98.9% (Wright et al., BMC Research Notes).
The anatomy ultrasound at 18 to 22 weeks can show the sex “if the fetus is in a good position,” as ACOG puts it. A review found sonographic sex assignment after 13 weeks accurate in 99 to 100% of babies with typical genitals (Odeh et al., 2009). The NHS still notes that a sonographer can’t be 100% certain.
For a timeline of when each becomes possible, see when can you find out the gender.
At-home DNA and urine kits: what the evidence covers
At-home kits are where the “100 percent accurate” promises usually live, and the evidence is thinnest.
Finger-prick blood kits use the same principle as cfDNA: they look for Y-chromosome DNA in the pregnant parent’s blood. The 2012 review noted that companies offer mail-order fetal sexing from a home finger-prick sample, but it didn’t evaluate them, and we found no independent, peer-reviewed study that does. The accuracy figures you’ll see come from the companies.
Two findings from the lab studies matter if you’re considering one:
- Timing. In the JAMA review, blood tests done before 7 weeks detected boys only 74.5% of the time, and the authors called them unreliable. A kit that promises an answer before 7 weeks is working in the range the review called unreliable.
- Handling. The 2012 review linked some wrong or failed results to samples not processed within 48 hours and to low amounts of fetal DNA. MedlinePlus Genetics adds that testing too early and sampling errors can leave too little fetal DNA for a result.
Urine kits that promise a color change for boy or girl have no good evidence. The same JAMA review concluded that “tests using urine” were unreliable.
If you use a home kit, treat the answer as provisional until a scan or a clinical test confirms it, especially if it says “girl”, since a girl result is the absence of a Y signal.
Heart rate, charts and old wives’ tales
None of the folk methods beat a coin toss in the studies that tested them.
- Heart rate. A 2023 meta-analysis of five studies and 4,308 fetuses found no meaningful difference in first-trimester heart rate between boys and girls (Nouri et al., BMC Pregnancy and Childbirth). Our post on baby heart rate and gender has the numbers.
- The Chinese chart. Applied to 2,840,755 Swedish births, it matched the real sex no better than chance (Villamor et al., 2010). More in our guide to the Chinese gender chart.
- Bump shape and morning sickness. In a Johns Hopkins study of 104 women, the shape of the bump and ordinary morning sickness weren’t related to the baby’s sex (Perry et al., Birth, 1999).
- Ramzi and skull theory. Ramzi’s method came from a 2011 article that its publisher now says “has not been confirmed in any additional studies” (Contemporary OB/GYN); skull theory has never been studied. See Ramzi and skull theory.
- Ring test, baking soda, cravings. We found no published study of any of them.
- The Drano test. Drain cleaners often contain sodium hydroxide, which can cause poisoning by touch, swallowing or breathing it in (MedlinePlus). Skip it. Call Poison Help at 1-800-222-1222 if anyone is exposed.
Each tale is rated in boy or girl old wives’ tales, and the early-scan method sonographers do use is covered in nub theory.
Why even the best tests miss sometimes
A few things make “100 percent” impossible, even for medical tests:
- Too little fetal DNA. Blood tests need enough placental DNA in the sample; MedlinePlus says the fetal fraction generally must be above 4 percent, which usually happens around 10 weeks.
- The view on a scan. If the legs are crossed or the baby faces away, the sonographer may not be able to tell.
- Genital development. Odeh’s team reported a 14-week baby who looked female on ultrasound and turned out to be a boy with hypospadias (Odeh et al., 2008).
- Human error. Labels, sample handling and reading the result all involve people.
If the sex matters for a medical reason, your provider will decide whether to confirm it with another test.
Guesses that are honest about being guesses
If you’re waiting for a real answer, or keeping it a surprise until birth, a guess can still be fun as long as everyone knows it’s a game. SeeMyBaby’s gender reveal party is built that way: you pull the ribbon on a gift box, and after you buy the gender add-on on the paywall, the box opens on a drumroll with pink or blue balloons and confetti. What it shows is a playful AI guess, just for fun, not a medical result. If you already know the sex, you tell the app and no guess is made.
Sources
- American College of Obstetricians and Gynecologists: Prenatal Genetic Diagnostic Tests, Prenatal Genetic Screening Tests and Ultrasound Exams
- Devaney SA, et al. Noninvasive fetal sex determination using cell-free fetal DNA: a systematic review and meta-analysis. JAMA, 2011.
- Wright CF, et al. Non-invasive prenatal diagnostic test accuracy for fetal sex using cell-free DNA: a review and meta-analysis. BMC Research Notes, 2012.
- MedlinePlus Genetics: What is noninvasive prenatal testing (NIPT)?
- Odeh M, et al. Sonographic fetal sex determination. Obstetrical & Gynecological Survey, 2009; and Hypospadias mimicking female genitalia, Journal of Clinical Ultrasound, 2008.
- NHS: Ultrasound scans in pregnancy
- Nouri S, et al. The role of fetal heart rate in first trimester sonograms in prediction of fetal sex. BMC Pregnancy and Childbirth, 2023.
- Villamor E, et al. Accuracy of the Chinese lunar calendar method to predict a baby’s sex. Paediatric and Perinatal Epidemiology, 2010.
- Perry DF, et al. Are women carrying “basketballs” really having boys? Birth, 1999.
- MedlinePlus: Sodium hydroxide poisoning
More gender guides are on the SeeMyBaby blog.
Questions parents ask
What is the most accurate gender test?
CVS and amniocentesis, which read the chromosomes directly, but they’re done only for medical reasons. For everyone else, a cfDNA blood test from 10 weeks and the 18–22 week anatomy scan are the most accurate options.
Are at-home gender blood tests accurate?
They work on the same principle as clinical cfDNA testing, but we found no independent peer-reviewed study of the home kits. Lab research shows blood tests before 7 weeks are unreliable.
Can a gender blood test be wrong?
Yes, occasionally. In large reviews, blood-based tests missed some boys and some girls, often because there was too little fetal DNA in the sample or the test was done too early.
Does any old wives’ tale work for boy or girl?
No. The ones that have been studied, including heart rate, bump shape and the Chinese chart, performed at chance.
Also available in: العربية · Deutsch · 日本語 · 한국어 · Português (Brasil)
