Detection of hepatitis C virus (HCV) RNA by a second generation (ver 2) HCV bDNA-probe method (bDNA-probe) was compared with detection by the first generation (ver 1) assay. The two assays were performed simultaneously with the same serum samples of HCV genotypes 1b, 2a, 2b, 3a, and 3b. The positive rates with ver 1 were 82% for HCV genotype 1b (type 1b), 57.6% for HCV genotype 2a (type 2a), 75.0% for HCV genotype 2b (type 2b), 55.6% for HCV genotype 3a (type 3a), and 93.8% for HCV genotype 3b (type 3b). The positive rates with ver 2 were 95.0% for type 1b, 93.9% for type 2a, 83.3% for type 2b, 100% for type 3a, and 93.8% for type 3b. With Fisher's exact test, the detection rate for type 2a was significantly higher (P = 0.001) with ver 2 than with ver 1. We obtained regression lines using the HCV counts measured by bDNA-probe on the y axis and the HCV counts obtained by an HCV reverse transcriptase (RT)-competitive polymerase chain reaction method (competitive PCR) on the x axis. The gradients for types 1b, 2a, and 3b were greater with ver 2 compared to ver 1. The gradients for types 2a and 3b were the highest: for type 2a, y = 0.135x + 0.6 with ver 1 and y = 0.248x + 0.1 with ver 2; for type 3b, y = 0.366x + 0.1 with ver 1 and y = 0.727x + 0.3 for ver 2. In addition, HCV-RNA counts for all the genotypes tested in this study were significantly higher with ver 2 than with ver 1. Hence, we conclude that ver 2 of the bDNA-probe measures HCV-RNA counts closer to those obtained with competitive PCR than the ver 1 assay.