AMD Software: Adrenalin Edition™ Application
Designed to provide you with a powerful yet easy-to-use interface where you can quickly access the latest software features, driver updates and much more.
Scalable Performance that Delivers More for Your AI Workloads.
Take your AI workloads further with the AMD Radeon™ AI PRO R9700 and AMD ROCm™. The combination of 32GB of VRAM and multi-GPU support creates a powerful foundation for large language models and other compute-heavy tasks, while delivering exceptional value in its class.
Unlock the Full Power of AMD Radeon™ AI PRO with AMD ROCm™
Ready to put the AMD Radeon™ AI PRO R9700 to work on your most demanding AI workloads? For a seamless start, explore our step-by-step guide to deploying AMD ROCm™ , OpenCL™, and PyTorch® environment on AMD Radeon AI PRO GPUs.
32GB VRAM for Larger AI Models
Optimal VRAM Buffer for Advanced Local AI
Meet the memory demands of modern LLMs and text-to-image models on your desktop with AMD Radeon™ AI PRO R9700 Graphics featuring 32GB of VRAM.
Typical VRAM Usage by Popular Models
LLM
28GB
27GB
Text to Image
24GB
17GB
See endnote RPW-496
Massive Gains for Massive Models
Up to 5× Faster with 32GB VRAM
AMD Radeon™ AI PRO R9700 (32GB)
GeForce RTX 5080 (16GB)
Performance (Tokens/Second) | See endnote RPW-495
Smarter Value for Your AI Projects
AMD Radeon™ AI PRO R9700 sets a new standard for cost-effective AI acceleration, delivering consistently higher value in high-VRAM workloads.
156%
148%
139%
177%
179%
158%
153%
170%
100%
AMD Radeon™ AI PRO R9700
Nvidia RTX 4500 (Blackwell)
Value (Performance / $), based on Average Tokens/Second | See endnotes RPW-502, 506
Our Trusted Partners
Discover Radeon™ AI PRO graphics cards manufactured by top AIB partners.
Shop from Your Regional Retailer
- North America
- Europe
- Asia Pacific
Model Specifications
- Radeon AI PRO R9000 Series
AMD Radeon AI PRO R9000 Series Graphics with AMD RDNA 4 Architecture
Radeon™ PRO W6000 and W7000 Series and Radeon™ AI PRO R9000 Series graphics cards (and later models) are not designed nor recommended for datacenter usage. Use in a datacenter setting may adversely affect manageability, efficiency, reliability, and/or performance. GD-239a.
Footnotes
RPW-495: Testing as of May 2025 by AMD. Average tokens per second of three runs, dropping edge cases where the model starts spiraling (more than 2k thinking tokens) to standardize response length. No speculative decode. All tests conducted on LM Studio 0.3.15 (Build 11). Vulkan Llama.cpp 1.28 used for AMD, NVIDIA-recommended CUDA 12 llama.cpp 1.30 with Flash Attention used for NVIDIA. Short Prompt: "How long would it take for a ball dropped from 10 meter height to hit the ground?“ Long Prompt: “Summarize the following in exactly five lines: [Insert Scene 1 Act 1 of Romeo and Juliet]”, Models tested: Phi 3.5 MoE Q4 K M, Mistral Small 3.1 24B Instruct 2503 Q8, DeepSeek R1 Distill Qwen 32B Q6, Qwen 32b Q6 System specifications: AMD Ryzen™ 9 7900X, 32GB DDR5 RAM 6000 MT/s, Windows 11 PRO 24H2, AMD Radeon™ AI PRO R9700 32GB using Adrenalin 25.6.1 RC vs AMD Ryzen™ 9 7900X, 32GB DDR5 RAM 6000 MT/s, Windows 11 PRO 24H2 with NVIDIA GeForce RTX 5080 and GeForce 576.4 drivers. Performance may vary. RPW-495.
RPW-496: Testing as of May 2025 by AMD using DeepSeek R1 Distill Qwen 32B Q6, Mistral Small 3.1 24B Instruct 2503 Q8, Flux.1 Schnell, SD 3.5 Medium models. Tested on a System with AMD Ryzen 9 7900X CPU, Radeon AI PRO R9700 GPU, 32GB DDR5 RAM, 1TB Storage, Windows 11 PRO 24H2, Adrenalin 25.6.1 RC drivers, ComfyUI - PyTorch 2.4 on Windows. System configurations may vary yielding different results. RPW-496
RPW-502: Testing as of September 2025 by AMD using DeepSeek R1 Distill Qwen 7B, DeepSeek R1 Distill Qwen 14B FP8-dynamic, Qwen3 14B FP8-dynamic, DeepSeek R1 Distill Llama 8B, Llama-3.1 8B-Instruct, DeepSeek R1 Distill Llama 8B FP8-dynamic, Llama-3.2 1B Instruct, Llama-3.2 3B-Instruct benchmarks. Tested on a system with ASUS WRX90E-SAGE-SE, AMD Ryzen™ Threadripper™ PRO 9985WX, 128GB DDR5-5600,1TB Storage, OS Ubuntu 24.04.3 LTS, kernel 6.14.0-27-generic, ROCm 6.4.2, AMD Radeon AI PRO R9700 vs. the same system with Nvidia RTX 4500 (Blackwell). System configurations may vary yielding different results. RPW-502
RPW-506: As of October 1, 2025, the AMD Radeon AI PRO R9700 at MSRP of $1299 USD offers the most VRAM (32GB) among the current generation of Workstation and Consumer video cards under $1500 USD. Sources: https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/ (Consumer) | https://www.cdw.com/ (Workstation). RPW-506
Footnotes
RPW-495: Testing as of May 2025 by AMD. Average tokens per second of three runs, dropping edge cases where the model starts spiraling (more than 2k thinking tokens) to standardize response length. No speculative decode. All tests conducted on LM Studio 0.3.15 (Build 11). Vulkan Llama.cpp 1.28 used for AMD, NVIDIA-recommended CUDA 12 llama.cpp 1.30 with Flash Attention used for NVIDIA. Short Prompt: "How long would it take for a ball dropped from 10 meter height to hit the ground?“ Long Prompt: “Summarize the following in exactly five lines: [Insert Scene 1 Act 1 of Romeo and Juliet]”, Models tested: Phi 3.5 MoE Q4 K M, Mistral Small 3.1 24B Instruct 2503 Q8, DeepSeek R1 Distill Qwen 32B Q6, Qwen 32b Q6 System specifications: AMD Ryzen™ 9 7900X, 32GB DDR5 RAM 6000 MT/s, Windows 11 PRO 24H2, AMD Radeon™ AI PRO R9700 32GB using Adrenalin 25.6.1 RC vs AMD Ryzen™ 9 7900X, 32GB DDR5 RAM 6000 MT/s, Windows 11 PRO 24H2 with NVIDIA GeForce RTX 5080 and GeForce 576.4 drivers. Performance may vary. RPW-495.
RPW-496: Testing as of May 2025 by AMD using DeepSeek R1 Distill Qwen 32B Q6, Mistral Small 3.1 24B Instruct 2503 Q8, Flux.1 Schnell, SD 3.5 Medium models. Tested on a System with AMD Ryzen 9 7900X CPU, Radeon AI PRO R9700 GPU, 32GB DDR5 RAM, 1TB Storage, Windows 11 PRO 24H2, Adrenalin 25.6.1 RC drivers, ComfyUI - PyTorch 2.4 on Windows. System configurations may vary yielding different results. RPW-496
RPW-502: Testing as of September 2025 by AMD using DeepSeek R1 Distill Qwen 7B, DeepSeek R1 Distill Qwen 14B FP8-dynamic, Qwen3 14B FP8-dynamic, DeepSeek R1 Distill Llama 8B, Llama-3.1 8B-Instruct, DeepSeek R1 Distill Llama 8B FP8-dynamic, Llama-3.2 1B Instruct, Llama-3.2 3B-Instruct benchmarks. Tested on a system with ASUS WRX90E-SAGE-SE, AMD Ryzen™ Threadripper™ PRO 9985WX, 128GB DDR5-5600,1TB Storage, OS Ubuntu 24.04.3 LTS, kernel 6.14.0-27-generic, ROCm 6.4.2, AMD Radeon AI PRO R9700 vs. the same system with Nvidia RTX 4500 (Blackwell). System configurations may vary yielding different results. RPW-502
RPW-506: As of October 1, 2025, the AMD Radeon AI PRO R9700 at MSRP of $1299 USD offers the most VRAM (32GB) among the current generation of Workstation and Consumer video cards under $1500 USD. Sources: https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/ (Consumer) | https://www.cdw.com/ (Workstation). RPW-506