AMD Radeon AI PRO R9700 32GB vs NVIDIA RTX 5090 for local LLMs
61 reports on the AMD Radeon AI PRO R9700 32GB and 161 on the NVIDIA RTX 5090
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the AMD Radeon AI PRO R9700 32GB has 29.8 generation tokens/s (3 runs); prompt processing 1287.0 tokens/s (one run); the NVIDIA RTX 5090 has 73.2 generation tokens/s (5 runs). Those are community reports whose engines and settings differ, so the gap is a rough sign, not a controlled measure of the two cards. Memory bandwidth, which caps how fast a card can write, is 2.8x higher on the NVIDIA RTX 5090 (640 GB/s on the AMD Radeon AI PRO R9700 32GB, 1,792 on the NVIDIA RTX 5090). Across 3 shared models with generation measurements on both cards, the AMD Radeon AI PRO R9700 32GB's median is higher on 0 and the NVIDIA RTX 5090's on 3.
- Memory
- 32GB
- Memory bandwidth
- 640 GB/s
- FP16 compute
- 191 TFLOPS
- Reports
- 61
- Memory
- 32GB
- Memory bandwidth
- 1,792 GB/s
- FP16 compute
- 419 TFLOPS
- Reports
- 161
Measured on both
Generation and prompt processing speeds from plain runs (one device, one request, no speculative decoding, the whole model in the device's memory) of the same model size at the same bit class. Each side shows the engines and reported context behind its median. Context can mean a configured window or prompt depth, so these are community comparisons with differing settings. How to read these
| Model | AMD Radeon AI PRO R9700 32GB | NVIDIA RTX 5090 |
|---|---|---|
| Qwen3.8 27B4-bit | 29.83 runsPP 1287.0 (one run)llama.cpp, 1 not stated2K context, 2 not stated | 73.25 runsllama.cpp, vLLM, 2 not stated120K to 256K context, 2 not stated |
| Qwen3.5 35B · 3B active4-bit | 149.2one runPP 3075.0 (one run)llama.cppcontext not stated | 194.0one runllama.cppcontext not stated |
| Qwen3.6 35B · 3B active4-bit | 61.4one runPP 7800.0 (one run)vLLM256K context | 215.1one runllama.cppcontext not stated |
Estimated by the calculator
The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape. Where a model was also measured above, the measurement wins.
| Model | AMD Radeon AI PRO R9700 32GB | NVIDIA RTX 5090 |
|---|---|---|
| Qwen3.8 27B | 16.0 t/sQ6_K | 44.9 t/sQ6_K |
| Qwen3.5 35B · 3B active | 58.2 t/sQ5_K_M | 163.0 t/sQ5_K_M |
| Qwen3.6 35B · 3B active | 58.2 t/sQ5_K_M | 163.0 t/sQ5_K_M |
| Qwen3.8 125B · 6B active | 18.7 t/sQ2_K, experts in RAM | 20.3 t/sQ2_K, experts in RAM |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 27B | 16.0 t/sQ6_K | 44.9 t/sQ6_K |
| Gemma 4 26B · 4B active | 40.9 t/sQ6_K | 114.5 t/sQ6_K |
| DeepSeek V4.1 Flash 552B · 16B active | doesn't fit at these settings | doesn't fit at these settings |
| GLM-5.3 320B · 18B active | doesn't fit at these settings | doesn't fit at these settings |
Change the context, RAM or model in the calculator: AMD Radeon AI PRO R9700 32GB · NVIDIA RTX 5090
Every report on each card, with its source: AMD Radeon AI PRO R9700 32GB · NVIDIA RTX 5090