Scientists have created immature sperm cells in a laboratory dish and injected them into eggs to produce mouse offspring.
The Chinese researchers say their stem cell technique could pave the way for exciting new treatments for male fertility.
British experts have called for the results to be independently verified and pointed out that any practical application is likely to be a long way off.
The mouse cells produced were technically "spermatids" - undeveloped sperm that lack tails and cannot swim.
Yet when they were injected into mouse eggs, mimicking a common IVF technique called Icsi (intracytoplasmic sperm injection), they delivered viable embryos and healthy, fertile babies.
In the UK, using spermatids in the same way to produce a pregnancy would be illegal.
Dr Jiahao Sha, from Nanjing Medical University, who co-led the research, reported in the peer-reviewed journal Cell Stem Cell, said: "If proven to be safe and effective in humans, our platform could potentially generate fully functional sperm for artificial insemination or in-vitro fertilisation techniques.
"Because currently available treatments do not work for many couples, we hope that our approach could substantially improve success rates for male infertility."
The scientists began with stem cells taken from mouse embryos which were exposed to a carefully mixed cocktail of chemicals.
This triggered their transformation into primordial germ cells, the first step on the developmental path to becoming sperm.
Next, the germ cells were exposed to testicular cells and testosterone in an attempt to mimic the natural environment of the testes.
When the resulting spermatids were injected into mouse eggs, they proved capable of producing embryos that developed normally.
Scientists have previously taken early steps in the process of creating artificial sperm and eggs in the laboratory.
In 2011 a Japanese team produced mouse germ cells from stem cells which eventually developed into healthy viable sperm, but only after they were injected into the testicles of male mice.
Infertility affects around 15% of couples and can be traced to the man in about a third of cases.
A major cause of male infertility is the failure of pre-cursor cells in the testes to undergo a special type of cell division called meiosis.
In 2014 a team of distinguished reproductive biologists writing in the journal Cell proposed a set of "gold standard" criteria to prove that all the essential steps of meiosis have taken place in artificially created eggs or sperm.
They included showing evidence of correct DNA content in the cell nucleus at specific meiotic stages, normal chromosome number and organisation, and the ability of the engineered cells to produce viable offspring.
The Chinese team claims to have passed all these tests.
Dr Sha said: "Our method fully complies with the gold standards recently proposed by a consensus panel of reproductive biologists, so we think that it holds tremendous promise for treating male infertility."
Scientists in the UK praised the "mammoth" achievement of their Chinese colleagues but said there were still many obstacles to be overcome before sperm-like cells grown in the laboratory could be of use to infertile men.
Professor Richard Sharpe, from the Medical Research Council Centre for Reproductive Health at the University of Edinburgh, said safety was a major issue.
"Bear in mind that if germ cells do not format their DNA correctly, it may not only affect the resulting individual but might also affect the next generation," he pointed out.
Allan Pacey, Professor of Andrology at the University of Sheffield, said the study was an "interesting step forward" but added: "It's important to note that the sperm-like cells produced in the study were not fully mature sperm as we might know them.
"In spite of these encouraging results, we are still some way from immediately applying this technique as a potential cure for human male infertility .. it remains to be seen if this technique could be applied in humans to create sperm-like cells that might be useable in IVF."
Professor Azim Surani, director of germline and epigenomics research at The Gurdon Institute, Cambridge University, said: "This research is potentially interesting but requires independent verification.
"There are some surprising elements, such as the highly accelerated rate of development of spermatid-like cells in culture.
"It also remains to be seen if this protocol could work with human stem cells because of the known differences between the mouse and human germline, and the protracted nature of human germline development."
But Professor Chris Barratt, from the University of Dundee, called the study a "landmark".
He added: "Importantly, although the efficiency of the process remains to be improved, the authors (using Icsi) achieve live births that are themselves fertile.
"Whilst human work is some way off, it is closer than we think."