APEX Versus keeps a live Power Rankings page — every entity that's ever been in a real, AI-scored comparison, ranked by their average score. It updates continuously as new comparisons get run, which means it's also a genuine, unfiltered record of what the AI actually thinks, not a curated list. As of this snapshot, the top of it looks like this:
| # | Entity | Score | Battles |
|---|---|---|---|
| 1 | Novak Djokovic | 9.9 | 2 |
| 2 | Michael Jordan | 9.7 | 1 |
| 3 | Son Goku | 9.7 | 5 |
| 4 | Goku | 9.6 | 1 |
| 5 | Katie Ledecky | 9.6 | 2 |
| 6 | Rafael Nadal | 9.6 | 1 |
| 7 | Simone Biles | 9.6 | 3 |
| 8 | The Dark Knight | 9.6 | 1 |
| 9 | Vegeta | 9.6 | 2 |
| 10 | Cristiano Ronaldo | 9.5 | 3 |
Full, always-current list: /power-rankings (top 50).
Six of the top ten are real, non-fictional athletes — not superheroes, not anime protagonists. A tennis player and an Olympic gymnast are outscoring Batman. That looks wrong at first glance if you assume "Power Rankings" means "strongest in an all-out fight." It doesn't, and that's the actually interesting finding here.
APEX Versus scores every entity on 7 dimensions specific to its own category (full breakdown on the Methodology page) — an athlete is judged on skill, athleticism, achievements, consistency, clutch factor, legacy, and versatility, out of 100 each, then averaged into a single 0-10 Apex Score. A fictional anime character is judged on an entirely different 7 dimensions — power, speed, durability, technique, battle IQ, willpower, potential.
The Power Rankings page puts every Apex Score, regardless of category, onto one shared 0-10 list. That's genuinely useful for browsing — but it means Djokovic's 9.9 and Goku's 9.7 aren't measuring the same thing. Djokovic's score says "about as dominant within tennis as it's possible to be." Goku's says "extremely powerful within his own anime power scale." Put them side by side on one list and it reads like a direct power comparison, when it's really two separate excellent-in-context scores landing near the same number.
The honest takeaway: the Power Rankings page is best read as "who's closest to the ceiling of their own category," not "who would beat whom." For an actual head-to-head verdict between two specific entities, that's what the real comparison tool is for — cross-category matchups get scored together, in the same call, specifically for that pairing.
Look closely at the list above and you'll notice "Son Goku" and "Goku" sitting at #3 and #4 as separate entries. Further down the full list, the same thing happens with "Naruto" and "Naruto Uzumaki," and with "The Dark Knight" and "Batman" — the same character, tracked as two different entities because different users typed the name differently.
This is a real, visible limitation of building rankings from free-text user input rather than a curated database: the system has no way to know "Goku" and "Son Goku" are the same person unless the exact name string matches. It's honest to say so rather than pretend the rankings are cleaner than they are — and it's also exactly the kind of thing that gets more accurate as more people search the same canonical name repeatedly.
Most of the top 20 are backed by just 1-3 real comparisons. Djokovic's #1 spot rests on 2 data points. This isn't a curated "greatest of all time" list — it's a live, growing dataset, and a single new comparison can meaningfully move an early entry's average. Worth remembering before treating any single ranking as definitive, and exactly why this is dated as a snapshot rather than presented as a permanent verdict.