AMD Instinct MI250X 128GB vs AMD Instinct MI325X for local LLMs
0 reports on the AMD Instinct MI250X 128GB and 0 on the AMD Instinct MI325X
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 1.8x higher on the AMD Instinct MI325X (3,277 GB/s on the AMD Instinct MI250X 128GB, 6,000 on the AMD Instinct MI325X). The estimates below compare them on the models people run most.
- Memory
- 128GB
- Memory bandwidth
- 3,277 GB/s
- FP16 compute
- 383 TFLOPS
- Reports
- 0
- Memory
- 256GB
- Memory bandwidth
- 6,000 GB/s
- FP16 compute
- 1307.4 TFLOPS
- Reports
- 0
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 | AMD Instinct MI250X 128GB | AMD Instinct MI325X |
|---|---|---|
| Qwen3.8 27B | 33.4 t/sFP16 | 61.1 t/sFP16 |
| Qwen3.8 125B · 6B active | 119.8 t/sQ6_K | 220.0 t/sQ8_0 |
| DeepSeek V4 Flash 284B · 13B active | 318.2 t/sQ2_K | 245.1 t/sQ5_K_M |
| Qwen3.6 35B · 3B active | 120.2 t/sFP16 | 220.0 t/sFP16 |
| Qwen3.6 27B | 33.4 t/sFP16 | 61.1 t/sFP16 |
| Gemma 4 26B · 4B active | 90.1 t/sFP16 | 165.0 t/sFP16 |
Change the context, RAM or model in the calculator: AMD Instinct MI250X 128GB · AMD Instinct MI325X
Every report on each card, with its source: AMD Instinct MI250X 128GB · AMD Instinct MI325X