AMD MI325X
Data center GPU · CDNA3
The MI325X is an evolution of the MI300X. It has better specs than the NVIDIA H200, but is generally available at lower prices.
Its large VRAM and high memory bandwidth make it especially useful for LLM inference applications. Software support for AMD GPUs has historically been less good than for equivalent NVIDIA GPUs, though this has been improving.
Tech specs
Cloud rental prices
1 provider · lowest price per GPU, per hour
Additional providers who may offer AMD MI325X without public pricing: TensorWave, DigitalOcean
Models that fit in VRAM
| Total VRAM | From | 16-bit inference | 8-bit inference | 4-bit inference | |
|---|---|---|---|---|---|
| 8× MI325X | 2,048 GB | $16.00/h | | | |
Open-weights models that fit in VRAM. Estimated using 🤗 accelerate, plus approximation for up to 8K context.
Media
Peak theoretical performance
Performance figures assume no sparsity; in cases where only sparse performance figures are published by the manufacturer, these are halved to give approximate dense performance.
Resources
Detailed documentation from the manufacturer.
Similar GPUs
Other accelerators you might compare.