I haven't been able to confirm nor refute this. However, being sloppy, it looks like pure cocaine has an LD50 of about 95 mg/kg [1] while caffeine appears to have an LD50 of about 367mg/kg [2]. These are numbers from rat and mouse studies, with different routes of administration. The cocaine LD50, for instance, was obtained by IP (intraperitoneal) injection, while the caffeine administration was oral. I have not been able to find any studies that give an LD50 for both substances in a single animal model with the same route of administration. Let's play with these numbers as they are.

The IC50 (inhibitory constant, a measure of sufficiency of plasma concentration to inhibit 50% of affected biological processes [how I'll estimate potency]) of caffeine in the human model is estimated at a concentration of about 67microMolar. Cocaine's IC50 appears to be about 4.4mM.[4]

The major difference in routes of administration will be how much of the substance ends up in plasma vs metabolites. Making the genuinely wild assumption that they have similar plasma effects from the same RoA, it would appear the IC50 of cocaine is about an order magnitude lower than that of caffeine (meaning you need less), while its LD50 is about a third that of caffeine (meaning less will kill you [but also keeping in mind that the LD50 was found via injections]) but still within the same order of magnitude approximation.

This to me indicates that cocaine is the preferable compound, even though I still think neither of them is great intransally. I will say, however, I absolutely prefer my cocaine as a drink and caffeine as a snifter. I have an intranasal cocktail of oxytocin, kanna, and caffeine that I'll bump once or twice a week. I quite like it, but people who rail lines of caffeine simply befuddle me.

[1] https://pubmed.ncbi.nlm.nih.gov/7178201/

[2] https://pubmed.ncbi.nlm.nih.gov/27461039/

[3] https://jnm.snmjournals.org/content/jnumed/early/2012/09/07/...

[4] https://pubmed.ncbi.nlm.nih.gov/11561083/