Every catalog hardware unit ranked by composite score (0–1000): measured tok/s, VRAM fit, ecosystem support, perf-per-watt. 2 of 154 ranks anchored to a measured benchmark — the rest are honestly flagged as extrapolated or estimated.
Methodology: /methodology · Run your own: curl -fsSL runlocalai.co/bench.mjs -o bench.mjs && node bench.mjs
40 units shown · sorted by score
| # | Hardware | Tier | Score | Data |
|---|---|---|---|---|
| 1 | AMD Instinct MI355X amd · workstation · 288GB | B | 626 | Estimated |
| 2 | AMD Instinct MI350X amd · workstation · 288GB | B | 626 | Estimated |
| 3 | AMD Instinct MI300X amd · workstation · 192GB | B | 621 | Estimated |
| 4 | AMD Instinct MI300A (APU) amd · workstation · 128GB | B | 620 | Estimated |
| 5 | AMD Instinct MI325X amd · workstation · 256GB | B | 615 | Estimated |
| 6 | AMD Instinct MI250X amd · workstation · 128GB | B | 614 | Estimated |
| 7 | AMD Instinct MI210 amd · workstation · 64GB | B | 587 | Estimated |
| 8 | AMD Radeon RX 7900 XTX amd · enthusiast · 24GB | C | 420 | Estimated |
| 9 | AMD Radeon RX 7900 XT amd · enthusiast · 20GB | C | 365 | Estimated |
| 10 | AMD Radeon RX 9060 XT amd · mid · 16GB | C | 339 | Estimated |
| 11 | AMD Radeon RX 9070 XT amd · high · 16GB | C | 332 | Estimated |
| 12 | AMD Radeon RX 9070 amd · high · 16GB | C | 332 | Estimated |
| 13 | AMD Radeon RX 7800 XT amd · high · 16GB | C | 329 | Estimated |
| 14 | AMD Radeon RX 7900 GRE amd · high · 16GB | C | 319 | Estimated |
| 15 | AMD Radeon RX 6950 XT amd · enthusiast · 16GB | C | 316 | Estimated |
| 16 | AMD Radeon RX 6800 amd · high · 16GB | C | 304 | Estimated |
| 17 | AMD Radeon RX 6800 XT amd · enthusiast · 16GB | C | 302 | Estimated |
| 18 | AMD Radeon RX 6900 XT amd · enthusiast · 16GB | C | 302 | Estimated |
| 19 | AMD Ryzen AI Max+ 395 (Strix Halo) amd · workstation | D | 288 | Estimated |
| 20 | AMD Radeon RX 9070 GRE amd · high · 12GB | D | 267 | Estimated |
| 21 | AMD Radeon RX 7700 XT amd · mid · 12GB | D | 265 | Estimated |
| 22 | AMD Radeon RX 6750 XT amd · high · 12GB | D | 265 | Estimated |
| 23 | AMD Radeon RX 7600 XT amd · mid · 16GB | D | 256 | Estimated |
| 24 | AMD Radeon RX 6700 XT amd · high · 12GB | D | 255 | Estimated |
| 25 | AMD Ryzen AI 9 HX 370 (Strix Point) amd · pc-npu | D | 246 | Estimated |
| 26 | AMD Radeon RX 5700 XT amd · high · 8GB | D | 214 | Estimated |
| 27 | AMD Radeon RX 6650 XT amd · mid · 8GB | D | 212 | Estimated |
| 28 | AMD Radeon RX 6600 XT amd · mid · 8GB | D | 208 | Estimated |
| 29 | AMD Radeon RX 6600 amd · mid · 8GB | D | 202 | Estimated |
| 30 | AMD Radeon RX 580 8GB amd · mid · 8GB | D | 172 | Estimated |
| 31 | AMD Radeon RX 5500 XT 8GB amd · entry · 8GB | D | 167 | Estimated |
| 32 | AMD EPYC 9005 (Zen 5, Turin) amd · workstation | D | 161 | Estimated |
| 33 | AMD Radeon RX 5600 XT amd · mid · 6GB | D | 158 | Estimated |
| 34 | HP ZBook Ultra G1a (Ryzen AI Max+ PRO 395) amd · workstation | D | 155 | Estimated |
| 35 | Framework Desktop (Ryzen AI Max+ 395) amd · workstation | D | 153 | Estimated |
| 36 | GMKtec EVO-X2 (Ryzen AI Max+ 395) amd · workstation | D | 153 | Estimated |
| 37 | Framework Laptop 16 (RX 7700S) amd · mid · 8GB | D | 147 | Estimated |
| 38 | AMD Radeon 880M (Strix Point iGPU) amd · entry | D | 132 | Estimated |
| 39 | AMD Radeon RX 570 amd · entry · 4GB | D | 131 | Estimated |
| 40 | AMD Radeon 780M (Phoenix iGPU) amd · entry | D | 127 | Estimated |
Amazon search links — we may earn a small commission at no extra cost to you. How we make money.
Steady-state tok/s on a representative 7B/8B Q4 model. Measured from real benchmark rows, or extrapolated from VRAM bandwidth × runtime-stack efficiency.
How comfortably the rig holds 7B / 32B / 70B class models. Apple unified memory counts; NPU/SoC system RAM counts.
CUDA / MLX / ROCm / Vulkan reach. Real-world friction the operator hits when installing tools.
Tok/s per watt. Mobile / NPU class scores well; dense desktop GPUs trade efficiency for absolute throughput.
A confidence multiplier (1.0 measured · 0.85 extrapolated · 0.7 estimated) discounts the headline so we don't pretend to know more than we do. Score is recomputed on every page load against the latest catalog + benchmark data — submit your own run with runlocalai-bench --submit --hardware your-rig to firm up the numbers.