GPU price index
Lowest price per GB
Lowest price per 100 GB/s
Lowest price per TOPS
Latest release
Biggest 24h drops
- 1GeForce RTX 5070 Ti$1,200-6.3%
- 2Tesla V100 32GB$720-2.6%
- 3GeForce RTX 5070$775-1.9%
- 4GeForce RTX 5090$6,500-1.5%
- 5Radeon RX 7800 XT$568-1.2%
Biggest 24h rises
- 1RTX PRO 4000 Blackwell$2,950+7.7%
- 2MacBook Pro M5 Max 128GB$7,250+6.6%
- 3Radeon RX 9070 XT$795+6.0%
- 4Mac Studio M3 Ultra 512GB$22,000+2.3%
- 5Tesla V100 16GB$305+1.7%
Most listings
- 1GeForce RTX 3090$1,55080 listings
- 2GeForce RTX 4090$3,19940 listings
- 3Tesla V100 16GB$30534 listings
- 4GeForce RTX 4080$1,26330 listings
- 5Radeon RX 7900 XTX$99925 listings
Latest releases: minimum VRAM
- 1
Hy4 preview770B18 GB - 2
GLM 5.3744B41 GB - 3GLM 5.3 Flash320B30 GB
- 4
Qwen3.8 Flash Next176B4 GB - 5
DeepSeek V4 Pro1,600B17 GB
Local LLM GPU Ranking
eBay
| # | Model | 30d trend | Ecosystem | Buy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 |  RTX PRO 6000 Blackwell | $16,9854 listings | 0.0% | 96 | 1,792 | 500 | 1,000 | 600 | $177 | $948 | $17 | CUDA | Check prices | |
| 2 | GeForce RTX 5090 | $6,50012 listings | -1.5% | 32 | 1,792 | 419 | 838 | 575 | $203 | $363 | $7.8 | CUDA | Check prices | |
| 3 | GeForce RTX 4090 | $3,19940 listings | 0.0% | 24 | 1,008 | 330.3 | 660.6 | 450 | $133 | $317 | $4.84 | CUDA | Check prices | |
| 4 | A100 40GB PCIe | $4,89913 listings | 0.0% | 40 | 1,555 | 312 | 624 | 250 | $122 | $315 | $7.9 | CUDA | Check prices | |
| 5 | GeForce RTX 5080 | $1,80014 listings | 0.0% | 16 | 960 | 225.1 | 450.2 | 360 | $113 | $188 | $4 | CUDA | Check prices | |
| 6 | RTX PRO 4500 Blackwell | $4,300Using previous price3 listings | 0.0% | 32 | 896 | 202.1 | 404.3 | 200 | $134 | $480 | $10.6 | CUDA | Check prices | |
| 7 | GeForce RTX 4080 | $1,26330 listings | +1.0% | 16 | 717 | 195 | 389.9 | 320 | $79 | $176 | $3.24 | CUDA | Check prices | |
| 8 |  Radeon RX 9070 XT | $79512 listings | +6.0% | 16 | 640 | 195 | 389 | 304 | $49.7 | $124 | $2.04 | ROCm | Check prices | |
| 9 | GeForce RTX 5070 Ti | $1,20010 listings | -6.3% | 16 | 896 | 175.8 | 351.5 | 300 | $75 | $134 | $3.41 | CUDA | Check prices | |
| 10 | RTX PRO 4000 Blackwell | $2,9503 listings | +7.7% | 24 | 672 | 161.3 | 322.5 | 145 | $123 | $439 | $9.2 | CUDA | Check prices | |
| 11 | GeForce RTX 3090 | $1,55080 listings | +0.3% | 24 | 936 | 142.3 | 284.7 | 350 | $65 | $166 | $5.5 | CUDA | Check prices | |
| 12 | GeForce RTX 5070 | $77523 listings | -1.9% | 12 | 672 | 123.5 | 247 | 250 | $65 | $115 | $3.14 | CUDA | Check prices | |
| 13 | GeForce RTX 2080 Ti 22GB | $5513 listings | 0.0% | 22 | 616 | — | 215.2 | 250 | $25.1 | $89 | $2.56 | CUDA | Check prices | |
| 14 | GeForce RTX 5060 Ti 16GB | $70012 listings | -0.7% | 16 | 448 | 94.9 | 189.8 | 180 | $43.8 | $156 | $3.69 | CUDA | Check prices | |
| 15 | Instinct MI210 | $5,04611 listings | 0.0% | 64 | 1,638 | 181 | 181 | 300 | $79 | $308 | $27.9 | ROCm | Check prices | |
| 16 | Radeon PRO W7900 | $3,495Using previous price3 listings | 0.0%30h | 48 | 864 | 123 | 123 | 295 | $73 | $405 | $28.4 | ROCm | Check prices | |
| 17 | Radeon RX 7900 XTX | $99925 listings | 0.0% | 24 | 960 | 123 | 123 | 355 | $41.6 | $104 | $8.1 | ROCm | Check prices | |
| 18 | Radeon RX 7900 XT | $70024 listings | +0.1% | 20 | 800 | 103 | 103 | 315 | $35 | $88 | $6.8 | ROCm | Check prices | |
| 19 | Radeon RX 6950 XT | $5009 listings | 0.0% | 16 | 576 | 47.3 | 94.6 | 335 | $31.3 | $87 | $5.3 | ROCm | Check prices | |
| 20 | Instinct MI100 | $9997 listings | 0.0% | 32 | 1,229 | 92.3 | 92.3 | 300 | $31.2 | $81 | $10.8 | ROCm | Check prices | |
| 21 | Radeon RX 7800 XT | $56824 listings | -1.2% | 16 | 624 | 74.7 | 74.7 | 263 | $35.5 | $91 | $7.6 | ROCm | Check prices | |
| 22 | Instinct MI50 32GB | $7687 listings | -0.1% | 32 | 1,024 | 26.5 | 53 | 300 | $24 | $75 | $14.5 | ROCm | Check prices | |
| 23 | Mac Studio M3 Ultra 512GB | $22,00011 listings | +2.3% | 512 | 819 | — | — | 480 | $43 | $2,686 | — | Metal | Check prices | |
| 24 | Mac Studio M4 Max 128GB | $6,5474 listings | 0.0% | 128 | 546 | — | — | 480 | $51 | $1,199 | — | Metal | Check prices | |
| 25 | MacBook Pro M4 Max 128GB | $5,5493 listings | 0.0% | 128 | 546 | — | — | 140 | $43.4 | $1,016 | — | Metal | Check prices | |
| 26 | MacBook Pro M5 Max 128GB | $7,2505 listings | +6.6% | 128 | 614 | — | — | 140 | $57 | $1,181 | — | Metal | Check prices | |
| 27 | Tesla V100 16GB | $30534 listings | +1.7% | 16 | 900 | 112 | — | 250 | $19.1 | $33.9 | — | CUDA | Check prices | |
| 28 | Tesla V100 32GB | $72023 listings | -2.6% | 32 | 900 | 112 | — | 250 | $22.5 | $80 | — | CUDA | Check prices |
Choose hardware for local LLM deployment
Find the right GPU for local LLMs. Compare VRAM, GPU prices and model memory requirements to choose hardware that fits your models and budget.
What is VRAM, and how much do you need?
VRAM is the fast memory on a graphics card. Local LLM inference needs space for model weights, the context (KV) cache and runtime buffers. Quantization can reduce weight memory, while longer prompts and more simultaneous users need additional memory. A model fitting in VRAM does not guarantee a particular speed.
Compare model VRAM requirementsWhat runs on 16GB of VRAM?Plan a local deployment
Choose a model and quantization, set the context length, and check both GPU memory and system RAM. CPU or expert offloading can make a large model fit with less VRAM, but it depends on RAM bandwidth and can be much slower. Check CUDA, ROCm or Metal support before buying hardware.
Local LLM hardware and deployment guidesRead the rankings and price data
Compare VRAM capacity first, then bandwidth, power and software support. Our prices are median asking prices, not completed sales; sample counts, observation dates and stale-price labels help you judge their usefulness. INT8 throughput alone is not an LLM benchmark.
Price sources and calculation methodsPrices are median asking prices, not sold prices · Methodology