AMD Instinct MI250 128GB vs AMD MI250X 128GB

The MI250X delivers 6% more FP16 throughput than the Instinct MI250 (383 vs 362 TFLOPS dense).

Instinct MI250: CDNA 2, 2021 MI250X: CDNA 2, 2021
FP16 dense lead
+6%
MI250X: 383 vs 362 TFLOPS
Memory
128 vs 128 GB
Same capacity
Bandwidth
3.3 TB/s vs 3.3 TB/s
MI250X moves data faster
TDP
500 W vs 560 W
Instinct MI250 draws 60 W less

Specifications

Instinct MI250
MI250X
Architecture
CDNA 2CDNA 2
Launch Year
20212021
Form Factor
OAMOAM
Memory
128 GB128 GB
Memory Bandwidth
3.3 TB/s3.3 TB/s
TDP
500 W560 W
Process Node
6nm6nm

Performance (TFLOPS)

Instinct MI250
MI250X
FP64
45 TFLOPS 48 TFLOPS
FP32
45 TFLOPS 48 TFLOPS
TF32
No verified data available No verified data available
BF16
362 TFLOPS 383 TFLOPS
FP16
362 TFLOPS 383 TFLOPS
FP8
No verified data available No verified data available
FP6
No verified data available No verified data available
FP4
No verified data available No verified data available
INT8
362 TOPS 383 TOPS

FLOPS by Precision

What actually differs

Both GPUs launched in 2021: the Instinct MI250 on CDNA 2 and the MI250X on CDNA 2.

Dense throughput for the Instinct MI250 against the MI250X: FP64 45 vs 48 TFLOPS, FP32 45 vs 48 TFLOPS, FP16 362 vs 383 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 128 GB against 128 GB, fed at 3.3 TB/s versus 3.3 TB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 500 W for the Instinct MI250 and 560 W for the MI250X. At FP32 that works out to 0.09 against 0.09 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Is the Instinct MI250 faster than the MI250X?

At FP16 precision the MI250X reaches 383 TFLOPS dense against 362 TFLOPS for the Instinct MI250. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the Instinct MI250 or the MI250X?

Both GPUs carry 128 GB of memory. Memory bandwidth is 3.3 TB/s for the Instinct MI250 and 3.3 TB/s for the MI250X.

How much power do the Instinct MI250 and the MI250X draw?

The Instinct MI250 is rated at 500 W TDP and the MI250X at 560 W. On FP32 throughput per watt, the Instinct MI250 is the more efficient part.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.