Is it worth upgrading?
A newer card winning every row of a comparison table does not make it worth buying. This works out how much faster it actually is at your resolution, what that improvement costs per percentage point, and whether a cheaper card would get you most of the way there.
Both cards have to be ones CompareYourGPU tracks. The database currently covers current and recent desktop cards, so if you are coming from something older it may not be listed yet. See what is covered.
Your card: RTX 3060 Ti
What should you upgrade to?
Every in-production card that is a real step up at 1440p, with what each one costs per percentage point gained. Pick a target card above for the full verdict on a specific pair.
The largest step that is still sensible value, not simply the fastest card.
RTX 5090
$1,999 · +244.8% at 1440p
massive
- $8.2 per percentage point gained
- +24 GB of memory
- Clears the memory shortfall at 1440p
The least you can spend and still feel the difference.
RX 9070
$549 · +89.7% at 1440p
major
- $6.1 per percentage point gained
- +8 GB of memory
- Clears the memory shortfall at 1440p
The most performance gained per dollar spent.
RX 9070 XT
$599 · +117.2% at 1440p
massive
- $5.1 per percentage point gained
- +8 GB of memory
- Clears the memory shortfall at 1440p
Everything worth upgrading to, ranked by value
In-production cards at least 25% faster than the RTX 3060 Ti at 1440p, ordered by cost per percentage point gained.
| Card | Gain | Price | Per 1% gained | VRAM change |
|---|---|---|---|---|
| RX 9070 XT | +117.2% | $599 | $5.1 | +8 GB |
| RTX 5070 Ti | +124.1% | $749 | $6 | +8 GB |
| RX 9070 | +89.7% | $549 | $6.1 | +8 GB |
| RTX 5080 | +155.2% | $999 | $6.4 | +8 GB |
| RTX 5070 | +75.9% | $549 | $7.2 | +4 GB |
| RTX 5090 | +244.8% | $1,999 | $8.2 | +24 GB |
Gains are CompareYourGPU’s estimated native index. Prices are launch MSRP unless a verified street price is on record, and the cost of an upgrade is the price of the new card: CompareYourGPU does not model what your current one might sell for.