Welcome to the forum and thanks for your question. I'll try to help. Thanks for an opportunity to clarify (again) a very common question on this forum. I'm taking the opportunity for a blog-like response that I can use in response to future questions.
The confusion about seroconversion time (window period) comes from three main sources.
First, failure to understand that not all HIV tests are the same. Older antibody tests took longer than current ones, but not all web sites or public health agencies have caught up with the difference. Also, testing often now includes direct tests for the virus, such as PCR for HIV DNA and tests for p24 antigen. For the combination of such a test plus antibody testing (e.g., the "4th generation", "duo", or "combi" test), the window period is only 4 weeks. By that time, all infected people will have a positive result on one or both components of the two tests.
Second, legal conservatism. Some agencies lean over backward to take no chance of telling someone they don't have HIV if they may be infected. Their legal departments advise going even further than scientifically necessary in their advice about window periods. That's where you may see advice about 6 months, even though all antibody test manufacturers claim their tests are 100% reliable by 3 months.
Third, regulatory issues. A company producing a new HIV test may validate it by studying blood from a few thousand persons, of whom some are known to have HIV and others are known not be infected. For those with HIV, the company's research team may decide to only study persons with infection known to be present 3 months or more. The test may in fact detect everyone infected for 6 weeks or even less. But since they have only studied those infected for 3 months or more, the US Food and Drug Administration (FDA) requires them to say a negative result may not be reliable until 3 months have passed.
For the antibody tests, the last issue is the most confusing. Everything known about the speed with which antibody develops in response to a new HIV infection, and the ability of the tests to detect that antibody, indicate that the tests will pick up virtually all new infections by 6 weeks, and certainly by 8 weeks. But because of the regulatory standards, the manufacturer may not claim 100% reliability until 3 months. And given the manufacturers' stance, it's easy to understand why physicians, clinics, health departments, or other agencies feel they need to stick with 3 months in their formal advice.
I hope this helps clarify these issues. For more information, see these threads as well (read them all -- the main information is in the follow-up comments).
Best regards-- HHH, MD