We treat '7B-class' as open-weight models with 4–12B parameters — 8 in our catalog. The table ranks every catalog GPU that fully fits at least one of them at Q4 / 8K context, by average estimated decode tok/s across the models it fits. Fit means weights + KV cache + 1 GB runtime overhead within 95% of VRAM — not weights alone. Methodology.
| # | GPU | VRAM (GB) | Bandwidth (GB/s) | MSRP | Fits | Avg Q4 tok/s | Buy |
|---|---|---|---|---|---|---|---|
| 1 | B200 192GB SXM | 192 | 8,000 | $40,000 | 8/8 | 737.5 est. | — |
| 2 | MI300X 192GB | 192 | 5,300 | $15,000 | 8/8 | 488.6 est. | — |
| 3 | H200 141GB SXM5 | 141 | 4,800 | $35,000 | 8/8 | 442.5 est. | — |
| 4 | H100 80GB SXM5 | 80 | 3,350 | $30,000 | 8/8 | 308.8 est. | — |
| 5 | Gaudi 3 128GB | 128 | 3,276 | $12,500 | 8/8 | 302.0 est. | — |
| 6 | RTX 5090 32GB | 32 | 1,792 | $1,999 | 8/8 | 186.3 incl. measured | Search Amazon (affiliate) |
| 7 | A100 80GB SXM4 | 80 | 2,039 | $15,000 | 8/8 | 185.9 incl. measured | — |
| 8 | RTX 4090 24GB | 24 | 1,008 | $1,599 | 8/8 | 116.1 incl. measured | GIGABYTE GeForce RTX 4090 Gaming OC 24GB (affiliate) · More listings |
| 9 | RTX 6000 Ada 48GB | 48 | 960 | $6,800 | 8/8 | 106.5 incl. measured | Search Amazon (affiliate) |
| 10 | RTX 5080 16GB | 16 | 960 | $999 | 8/8 | 105.8 incl. measured | NVIDIA GeForce RTX 5080 Founders Edition (affiliate) · More listings |
| 11 | RTX 5070 Ti 16GB | 16 | 896 | $749 | 8/8 | 100.3 incl. measured | Search Amazon (affiliate) |
| 12 | RTX 3090 Ti 24GB | 24 | 1,008 | $1,999 | 8/8 | 92.9 est. | Search Amazon (affiliate) |
| 13 | RTX 4090D 24GB (China) | 24 | 1,008 | $1,599 | 8/8 | 92.9 est. | — |
| 14 | RTX 3090 24GB | 24 | 936 | $1,499 | 8/8 | 86.9 incl. measured | MSI GeForce RTX 3090 Ventus 3X 24G OC (affiliate) · More listings |
| 15 | RX 7900 XTX 24GB | 24 | 960 | $999 | 8/8 | 85.0 incl. measured | ASRock Radeon RX 7900 XTX Phantom Gaming 24GB OC (affiliate) · More listings |
| 16 | L40S 48GB | 48 | 864 | $8,000 | 8/8 | 79.7 est. | Search Amazon (affiliate) |
| 17 | Mac Studio M3 Ultra 96GB | 96 | 819 | $3,999 | 8/8 | 75.5 est. | Search Amazon (affiliate) |
| 18 | RTX 5070 12GB | 12 | 672 | $549 | 8/8 | 75.5 incl. measured | Search Amazon (affiliate) |
| 19 | RX 7900 XT 20GB | 20 | 800 | $899 | 8/8 | 73.8 est. | Search Amazon (affiliate) |
| 20 | RTX 3080 10GB | 10 | 760 | $699 | 8/8 | 71.2 incl. measured | Search Amazon (affiliate) |
| 21 | RTX A6000 48GB | 48 | 768 | $4,500 | 8/8 | 70.8 est. | Search Amazon (affiliate) |
| 22 | RX 9070 XT 16GB | 16 | 640 | $599 | 8/8 | 69.6 incl. measured | Search Amazon (affiliate) |
| 23 | RX 9070 16GB | 16 | 640 | $549 | 8/8 | 69.2 incl. measured | Search Amazon (affiliate) |
| 24 | Mac Studio M2 Ultra 192GB | 192 | 800 | $6,999 | 8/8 | 68.9 incl. measured | Search Amazon (affiliate) |
| 25 | RTX 4080 Super 16GB | 16 | 736 | $999 | 8/8 | 67.8 incl. measured | ASUS ProArt GeForce RTX 4080 Super OC (affiliate) · More listings |
| 26 | RTX 4070 Ti Super 16GB | 16 | 672 | $799 | 8/8 | 67.2 incl. measured | Search Amazon (affiliate) |
| 27 | RX 7800 XT 16GB | 16 | 624 | $499 | 8/8 | 59.8 incl. measured | Search Amazon (affiliate) |
| 28 | RX 7900 GRE 16GB | 16 | 576 | $549 | 8/8 | 53.1 est. | Search Amazon (affiliate) |
| 29 | RTX 5060 Ti 16GB | 16 | 448 | $429 | 8/8 | 52.6 incl. measured | Search Amazon (affiliate) |
| 30 | Arc A770 16GB | 16 | 560 | $349 | 8/8 | 51.6 est. | Search Amazon (affiliate) |
| 31 | RTX 4070 Ti 12GB | 12 | 554 | $799 | 8/8 | 51.1 est. | Search Amazon (affiliate) |
| 32 | MacBook Pro M4 Max 128GB | 128 | 546 | $4,399 | 8/8 | 50.3 est. | Search Amazon (affiliate) |
| 33 | RTX 4070 12GB | 12 | 504 | $599 | 8/8 | 48.0 incl. measured | ASUS Dual GeForce RTX 4070 OC 12GB (affiliate) · More listings |
| 34 | RTX 4070 Super 12GB | 12 | 504 | $599 | 8/8 | 46.5 est. | Search Amazon (affiliate) |
| 35 | Arc B580 12GB | 12 | 456 | $249 | 8/8 | 42.0 est. | Search Amazon (affiliate) |
| 36 | RTX 3070 8GB | 8 | 448 | $499 | 8/8 | 41.3 est. | Search Amazon (affiliate) |
| 37 | RTX 5060 8GB | 8 | 448 | $299 | 8/8 | 41.3 est. | Search Amazon (affiliate) |
| 38 | RTX 5060 Ti 8GB | 8 | 448 | $379 | 8/8 | 41.3 est. | Search Amazon (affiliate) |
| 39 | RX 7700 XT 12GB | 12 | 432 | $449 | 8/8 | 39.8 est. | Search Amazon (affiliate) |
| 40 | MacBook Pro M3 Max 128GB | 128 | 400 | $4,199 | 8/8 | 37.6 incl. measured | Search Amazon (affiliate) |
| 41 | Arc B570 10GB | 10 | 380 | $219 | 8/8 | 35.0 est. | Search Amazon (affiliate) |
| 42 | RTX 3060 12GB | 12 | 360 | $329 | 8/8 | 34.8 incl. measured | MSI GeForce RTX 3060 Ventus 2X 12G OC (affiliate) · More listings |
| 43 | RTX 4060 Ti 16GB | 16 | 288 | $499 | 8/8 | 29.9 incl. measured | Search Amazon (affiliate) |
| 44 | T4 16GB | 16 | 320 | $2,000 | 8/8 | 29.5 est. | Search Amazon (affiliate) |
| 45 | RTX 4060 Ti 8GB | 8 | 288 | $399 | 8/8 | 26.6 est. | Search Amazon (affiliate) |
| 46 | RX 7600 8GB | 8 | 288 | $269 | 8/8 | 26.6 est. | Search Amazon (affiliate) |
| 47 | RX 7600 XT 16GB | 16 | 288 | $329 | 8/8 | 26.6 est. | Search Amazon (affiliate) |
| 48 | Mac mini M4 Pro 64GB | 64 | 273 | $2,199 | 8/8 | 25.2 est. | Search Amazon (affiliate) |
| 49 | RTX 4060 8GB | 8 | 272 | $299 | 8/8 | 25.1 est. | Search Amazon (affiliate) |
| 50 | Mac mini M4 32GB | 32 | 120 | $999 | 8/8 | 12.5 incl. measured | Search Amazon (affiliate) |