NVIDIA RTX 4000 Ada vs AMD Strix Halo 128GB for local LLMs
0 reports on the NVIDIA RTX 4000 Ada and 120 on the AMD Strix Halo 128GB
Which is faster for local LLMs?
Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 41% higher on the NVIDIA RTX 4000 Ada (360 GB/s on the NVIDIA RTX 4000 Ada, 256 on the AMD Strix Halo 128GB). The estimates below compare them on the models people run most.
- Memory
- 20GB
- Memory bandwidth
- 360 GB/s
- FP16 compute
- 106.9 TFLOPS
- Reports
- 0
- Memory
- 128GB unified
- Memory bandwidth
- 256 GB/s
- FP16 compute
- 59 TFLOPS
- Reports
- 120
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.
| Model | NVIDIA RTX 4000 Ada | AMD Strix Halo 128GB |
|---|---|---|
| Qwen3.8 27B | 16.8 t/sQ3_K_M | 3.5 t/sFP16 |
| Qwen3.8 125B · 6B active | 17.0 t/sQ2_K, experts in RAM | 12.6 t/sQ6_K |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | 33.5 t/sQ2_K |
| Qwen3.6 35B · 3B active | 66.6 t/sQ3_K_S | 12.6 t/sFP16 |
| Qwen3.6 27B | 16.8 t/sQ3_K_M | 3.5 t/sFP16 |
| Gemma 4 26B · 4B active | 37.7 t/sQ4_K_S | 9.5 t/sFP16 |
Change the context, RAM or model in the calculator: NVIDIA RTX 4000 Ada · AMD Strix Halo 128GB
Every report on each card, with its source: NVIDIA RTX 4000 Ada · AMD Strix Halo 128GB