What a talent value actually is
Every Pal carries three hidden whole numbers, one each for Health, Attack and Defense. They are always called IVs by players, but the game's own term is potentials or talent, and the official wiki documents how they convert:
bonus% = talent × 0.3
So talent 0 is a 0% bonus, talent 50 is +15%, and talent 100 is +30%. That 30% is the complete spread between the best and worst possible Pal of any given species — there is no hidden higher tier, and there are no letter grades. Any S/A/B ranking you have seen comes from a third-party tool applying its own cut-offs, not from the game.
The bonus applies to level gains, not base stats
This is the part most people get wrong, and it changes how much talent is worth. The percentage does not multiply the species' base stat; it applies to the stat the Pal gains as it levels. The practical result is a curve:
- At level 1 the difference between a 0 and 100 talent is almost nothing.
- By level 30 it is already around an 11% swing.
- By level 50, a good attack talent against a bad one is roughly a 12% damage difference — comparable to adding a third damage passive, and larger than a fourth.
That is why talent hunting is a mid-game concern rather than an early one. On a level 10 worker Pal it is noise; on a level 60 raid Pal it is the difference between two builds.
How to see and raise talent
Ability Glasses display the raw 0–100 numbers, and they are late-tier: technology level 34, ancient technology points, and Ancient Civilization Cores in the recipe. Most players breed blind for a long time before they can read the values at all.
To raise a specific stat, the matching fruit adds +10 per use, capped at 100: Attack Fruit for Attack, HP Fruit for Health, Defense Fruit for Defense. The calculator converts a target into a fruit count, which is the honest way to cost an upgrade — ten talent points per fruit, nothing more.
Because the cap is 100, a Pal sitting at 60 needs four fruit for that stat, and one at 95 needs a single fruit. Those last few points are where fruit are actually worth spending, since a 0-to-40 jump costs the same as a 95-to-100 one.
Inheritance: why perfect parents are not enough
Breeding does not copy talents straightforwardly. The game stores how many talent values an egg inherits from its parents in Combi_TalentInheritNum, and the value is 1, 2 or 3 — an egg takes at most three of the six talent values its two parents carry across the three stats, and anything not inherited is rolled fresh.
Two consequences follow, and both are easy to get wrong:
- At least one talent always transfers, which is why keeping high-talent parents is worth doing.
- No stat can be locked in. There is no equivalent of a guaranteed passive, so "perfect IV" breeding is a volume exercise: hatch many, glass the results, keep the best, repeat.
Should you care? For a combat Pal, Attack talent plus percentage Attack passives is the biggest single lever. For a base worker, Health and work speed matter more than Attack. For a Pal you will retire at level 20, talent is close to irrelevant — check it and move on.
Sources and standards
Where these figures come from
Every constant and formula on this page was taken from the published reference below and re-checked before it was published.
- Official Palworld Wiki — the talent to percentage conversion, the 0-100 range, the +30% ceiling and the fruit mechanics.
- Official Palworld Wiki — breeding mechanics referenced for the inheritance behaviour described here.
- Palworld Breeding — the Combi_TalentInheritNum inheritance count of 1-3 and the Ability Glasses requirements, checked against patch 1.0.3.
Talent to bonus reference
| Talent number | Bonus to level gains | Fruit to reach 100 |
|---|---|---|
| 0 | +0% | 10 |
| 10 | +3% | 9 |
| 20 | +6% | 8 |
| 30 | +9% | 7 |
| 40 | +12% | 6 |
| 50 | +15% | 5 |
| 60 | +18% | 4 |
| 70 | +21% | 3 |
| 80 | +24% | 2 |
| 90 | +27% | 1 |
| 100 | +30% | — |
Bonus is talent × 0.3. The fruit column is what the last stretch actually costs: a Pal at talent 0 needs ten fruit to reach the cap, and one at 90 needs a single fruit. Fruit cost the same regardless of where they are spent, so the last few points are always the cheapest talent you will ever buy.
Worked examples
A raid Pal at level 50
Attack talent 80 gives a +24% bonus to level gains, against +9% for a talent 30 Pal of the same species. The gap is comparable to adding a strong damage passive — and unlike a passive, it cannot be replaced later without rebreeding.
A base worker you will keep at level 20
At talent 50 a worker gets +15% on level gains, but at level 20 the absolute numbers are small enough that the difference between a good and a bad worker is a few points of work speed. Check the talent, note it, and do not rebreed for it.
Breeding for a perfect child
Two parents at talent 100 in all three stats carry six perfect values between them. An egg inherits at most three of those and rolls the rest fresh, so the chance of a perfect child is not a guarantee and there is no way to force it. The lever you actually control is how many eggs you hatch.