This review is the second in a series of four papers emanating from a workshop entitled “Developmental Toxicology—New Directions,” which was sponsored by the ILSI Health and Environmental Sciences Institute's (HESI) Developmental and Reproductive Toxicology Technical Committee. The present review analyzes the strengths and weaknesses of current developmental safety testing approaches in an effort to identify those strengths that should be retained in the future versus the weaknesses that should be eliminated. Workshop participants considered the following to be key strengths of current testing approaches: the integrated biology of pregnant animal models including pharmacokinetic and pharmacodynamic processes, the ability to detect low incidence malformations as well as maternally mediated toxicity, and the long history of use coupled with extensive historical data. A number of weaknesses were related to the resource-intensive nature of developmental toxicity testing (e.g., large number of animals, high costs, low throughput, the inability to keep pace with the demand for more toxicity data). Other weaknesses included the use of very high dose levels that often far exceed human exposure levels, the confounding influence of maternal toxicity, sparse understanding of basic developmental mechanisms and genetics of standard animal models relative to mouse or lower organisms, difficulties interpreting low incidence findings, and issues surrounding the interpretation of minor skeletal variations. An appreciation of these strengths and weaknesses is critical for the design of new approaches to developmental toxicity testing in the 21st century. Birth Defects Res (Part B) 92:395–403, 2011. © 2011 Wiley Periodicals, Inc.