Compute-балл — для general-purpose GPU compute: CUDA/OpenCL workload, научные расчёты, симуляции, видеомонтаж и энкодинг через NVENC. Формула: FP32 + (CUDA cores ÷ 1000) × 0.4 + memory bandwidth ÷ 100.
Фильтры: VRAM 10-13 ГБ + Consumer (для дома)
| # | Модель | Цена ₽ | Вычисления |
|---|---|---|---|
| #1 | NVIDIA GeForce RTX 4070 Ti Ada Lovelace | 102 200 | 27.7 |
| #2 | NVIDIA GeForce RTX 4070 SUPER Ada Lovelace | 63 810 | 26.3 |
| #3 | NVIDIA GeForce RTX 3080 12 GB Ampere | 41 000 | 24.9 |
| #4 | NVIDIA GeForce RTX 3080 Ti Ampere | 58 533 | 24.4 |
| #5 | NVIDIA GeForce RTX 5070 Blackwell 2.0 | 55 600 | 24.2 |
| #6 | AMD Radeon RX 9070 GRE RDNA 4.0 | 65 438 | 24.1 |
| #7 | AMD Radeon RX 7700 XT RDNA 3.0 | 130 652 | 23.5 |
| #8 | NVIDIA GeForce RTX 4070 AD103 Ada Lovelace | 49 990 | 22.1 |
| #9 | NVIDIA GeForce RTX 4070 GDDR6 Ada Lovelace | 49 990 | 22.0 |
| #10 | NVIDIA GeForce RTX 3080 Ampere | 33 000 | 21.0 |
| #11 | NVIDIA GeForce RTX 4070 Ada Lovelace | 49 990 | 21.0 |
| #12 | Intel Arc B580 Xe2-HPG | 24 436 | 11.7 |
| #13 | AMD Radeon RX 6750 XT RDNA 2.0 | 25 500 | 10.7 |
| #14 | AMD Radeon RX 6850M XT RDNA 2.0 | — | 10.7 |
| #15 | AMD Radeon RX 6750 GRE 12 GB RDNA 2.0 | — | 10.7 |
| #16 | Intel Arc A730M Xe-HPG | — | 9.9 |
| #17 | Intel Arc B570 Xe2-HPG | 21 214 | 9.8 |
| #18 | AMD Radeon RX 6700 XT RDNA 2.0 | 73 577 | 9.3 |
| #19 | NVIDIA GeForce RTX 3060 12 GB Ampere | 26 531 | 9.2 |
| #20 | NVIDIA GeForce RTX 3060 12 GB GA104 Ampere | 26 531 | 9.2 |
| #21 | NVIDIA GeForce RTX 2080 Ti Turing | 32 973 | 9.1 |
| #22 | AMD Radeon RX 6750 GRE 10 GB RDNA 2.0 | 23 929 | 8.9 |
| #23 | AMD Radeon RX 6700 RDNA 2.0 | 27 502 | 7.9 |
| #24 | NVIDIA GeForce GTX 1080 Ti Pascal | — | 5.9 |
| #25 | NVIDIA GeForce RTX 2060 12 GB Turing | 25 269 | 5.9 |
| #26 | NVIDIA GeForce GTX TITAN X Maxwell 2.0 | — | 2.7 |
Подробности расчёта — страница методики.