Very little informational content in this article. 90% of it is structured around building up the case that "Exxon knew exactly how much warming would occur" but there is very little information about the models themselves that appear to be the key piece of evidence. The one sentence that does mention them says "From 63% to 83% of those projections fit strict standards for accuracy and generally predicted correctly that the globe would warm about .36 degree (.2 degree Celsius) a decade." So we have an uncertain number of projections that match the outcome to within whatever amount of error "strict standards for accuracy" means. If the number is on the low end of 63% and the error is loosey-goosey then I can see why Exxon would say that their models may not be reliable since they would be getting 37% of their predictions wrong and some amount of the "correct" predictions would be off as well. People love to talk about broken clocks, I think we should be talking about broken models.