Diagnostic information allows an agent to predict the state of nature about the success of an investment project better than the prior. We analyze the optimal pricing scheme for selling diagnostic information to buyers with different, privately known, ex ante success probability. Investment costs and returns of successful projects are assumed to be the same for all buyers. The value of diagnostic information is the difference in expected payoffs with and without it, and we show that the willingness to pay for diagnostic information is nonmonotonic in the ex ante success probability. When the information seller can offer only one quality level, and negative payments are not allowed, we find that the optimal menu of (linear) contracts is remarkably simple. A pure royalty is offered to buyers with low ex ante success probability, and a pure fixed fee is offered to buyers with high ex ante success probability.