AMD Radeon Instinct MI300X vs NVIDIA RTX PRO 4000 Blackwell SFF

Comparison of AMD Radeon Instinct MI300X with 192 GB HBM3 and 19,456 cores vs NVIDIA RTX PRO 4000 Blackwell SFF with 24 GB GDDR7 and 8,960 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI300X outperforms NVIDIA RTX PRO 4000 Blackwell SFF by 609.12% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

AMD Radeon Instinct MI300X

94.1

AMD Radeon Instinct MI300X

94.1
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA RTX PRO 4000 Blackwell SFF

13.3

NVIDIA RTX PRO 4000 Blackwell SFF

13.3

Contents:

Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks Additional

Memory

Memory Size

🔥 +700% 192 ГБ
24 GB

Memory Type

HBM3 GDDR7

Memory Bandwidth

🔥 10.3 TB/s
432.0 GB/s

Memory Bus Width

8,192 бит 192 бит

ML Performance

FP16 (Half Precision)

🔥 +2,448% 653.7 TFLOPS
25.66 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +218% 81.72 TFLOPS
25.66 TFLOPS

FP64 (Double Precision)

🔥 +20,279% 81.72 TFLOPS
0.401 TFLOPS

CUDA Cores

🔥 +117% 19,456
8,960

RT Cores

No
🔥 70

Architecture & Compatibility

GPU Architecture

CDNA 3.0 Blackwell 2.0

SM (Streaming Multiprocessor)

No
🔥 70

PCIe Version

PCIe 5.0 x16 PCIe 5.0 x8

ML Software Support

CUDA Version

No 12.0

Clocks & Performance

Base Clock

🔥 +73% 1,000
577

Boost Clock

🔥 +47% 2,100
1,432

Memory Clock

🔥 +124% 2,525
1,125

Power Consumption

Recommended PSU

1150 W
🔥 -78% 250 W

Power Connector

None None

TDP/TGP

750 W
🔥 -91% 70 W

Rendering

Texture Units (TMU)

🔥 +334% 1,216
280

ROP

No
🔥 70

L2 Cache

16 MB
🔥 +200% 48 MB

Benchmarks

MLPerf, llama2-70b-99.9 (UNSET)

1 983 tokens/s

MLPerf, llama2-70b-99.9 (fp16)

1 740 tokens/s

MLPerf, llama2-70b-99.9 (fp8)

1 057 tokens/s

MLPerf, llama3.1-405b (UNSET)

30.4 tokens/s

MLPerf, llama3.1-405b (fp16)

34.8 tokens/s

llama.cpp, llama-2-7b-Q4_0

232.9 tokens/s

MLPerf, mixtral-8x7b (fp8)

5 975 tokens/s

Additional

Slots

OAM Module Dual-slot

Release Date

Dec. 6, 2023 Aug. 11, 2025

Display Outputs

No outputs
4x mini-DisplayPort 2.1b

Renting is cheaper than buying