AMD Instinct MI355X OAM vs AMD Instinct MI350X OAM

The Instinct MI355X delivers 9% more FP8 throughput than the Instinct MI350X (5,033 vs 4,614 TFLOPS dense).

Instinct MI355X: CDNA 4, 2025 Instinct MI350X: CDNA 4, 2025
FP8 dense lead
+9%
Instinct MI355X: 5,033 vs 4,614 TFLOPS
VRAM
288 vs 288 GB
Same capacity
Bandwidth
8.0 TB/s vs 8.0 TB/s
Even
TDP
1.4 kW vs 1.0 kW
Instinct MI350X draws 400 W less

Specifications

Instinct MI355X
Instinct MI350X
Architecture
CDNA 4CDNA 4
Launch Year
20252025
Form Factor
OAMOAM
VRAM
288 GB288 GB
Memory Bandwidth
8.0 TB/s8.0 TB/s
TDP
1.4 kW1.0 kW
Process Node
3nm/6nm3nm/6nm

Performance (TFLOPS)

Instinct MI355X
Instinct MI350X
FP64
78.6 TFlops 72.1 TFlops
FP32
157.3 TFlops 144.2 TFlops
FP16
2516.6 TFlops
5033.2 TFLOPS sparse
2309.6 TFlops
4619.2 TFLOPS sparse
BF16
2516.6 TFlops
5033.2 TFLOPS sparse
2309.6 TFlops
4619.2 TFLOPS sparse
FP8
5033.2 TFlops
10066.4 TFLOPS sparse
4614.0 TFlops
9227.4 TFLOPS sparse
INT8
5033.2 TFlops
10066.4 TOPS sparse
4614.0 TFlops
9227.4 TOPS sparse

FLOPS by Precision

What actually differs

Both GPUs launched in 2025: the Instinct MI355X on CDNA 4 and the Instinct MI350X on CDNA 4.

Dense throughput for the Instinct MI355X against the Instinct MI350X: FP64 79 vs 72 TFLOPS, FP32 157 vs 144 TFLOPS, FP16 2,517 vs 2,310 TFLOPS, FP8 5,033 vs 4,614 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 288 GB against 288 GB, fed at 8.0 TB/s versus 8.0 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 1.4 kW for the Instinct MI355X and 1.0 kW for the Instinct MI350X. At FP32 that works out to 0.11 against 0.14 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 MI355X faster than the Instinct MI350X?

At FP8 precision the Instinct MI355X reaches 5,033 TFLOPS dense against 4,614 TFLOPS for the Instinct MI350X. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the Instinct MI355X or the Instinct MI350X?

Both GPUs carry 288 GB of VRAM. Memory bandwidth is 8.0 TB/s for the Instinct MI355X and 8.0 TB/s for the Instinct MI350X.

How much power do the Instinct MI355X and the Instinct MI350X draw?

The Instinct MI355X is rated at 1.4 kW TDP and the Instinct MI350X at 1.0 kW. On FP32 throughput per watt, the Instinct MI350X is the more efficient part.

Can I rent the Instinct MI355X or the Instinct MI350X in the cloud?

Yes. Live cloud listings tracked by Flopper start at $6.16 per GPU hour. The rental pricing section on this page lists current providers and rates for both GPUs.

Where to Rent

AMD Instinct MI350X OAM

ProviderConfigurationPrice/GPU-hrChecked
DigitalOcean
1× Instinct MI350X OAM $6.16 4h ago View →
DigitalOcean
8× Instinct MI350X OAM $6.16 4h ago View →
All Instinct MI350X listings and price history →

Get Comparison Updates

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