Several genetically engineered mouse models are currently being examined for potential use in cancer hazard identification. We have undertaken an interlaboratory comparison of the performance of the CB6F1-Tg rasH2 transgenic mouse in cancer bioassays concurrently conducted in the United States and Japan. Chemicals selected for study included known human carcinogens (melphalan and cyclosporin A) and known rodent carcinogens (p-cresidine and vinyl carbamate) tested at carcinogenic doses, and non-carcinogens (p-anisidine and resorcinol) tested at appropriate high doses. Because of abdominal adhesions caused by the intraperitoneal dosing vehicle, melphalan was excluded from the study results. The remaining five studies showed similar results between the two laboratories conducting each study. Vinyl carbamate gave the strongest positive response inducing lung adenomas and carcinomas and splenic hemangiosarcomas. p-Cresidine was considered positive for urinary bladder transitional neoplasia. Cyclosporin A, p-anisidine, and resorcinol were negative in all studies. Although only five chemicals were successfully tested in this interlaboratory comparison, there was good concordance in outcome for the strong carcinogens and for the non-carcinogens. Successful testing of chemicals with less carcinogenic potential may require modifications in study design to include more animals and longer study duration.