The MI300A is a revolutionary APU combining Zen 4 CPU cores with CDNA 3 GPU compute on a single package. Its unified memory architecture eliminates data transfer overhead between CPU and GPU.
VRAM
128 GB
Memory
HBM3
Bandwidth
5300 GB/s
TDP
760W
Large Language Models
Training and inference for models like GPT-4, Llama 70B+
Deep Learning Training
High-performance training for neural networks
High-Throughput Inference
Optimized for batched inference workloads
Enterprise Deployment
Designed for 24/7 datacenter operations
APU combining Zen 4 CPU cores with CDNA 3 GPU. Unified memory architecture.
Estimates based on INT8 quantization. Actual fit depends on framework and batch size.
Added Jan 27, 2026
Last updated: Jan 27, 2026
Explore models, compare pricing and benchmarks, and right-size your infrastructure — all in one place.