NVIDIA RTX 5090 Laptop 24GB vs AMD Strix Halo 128GB for local LLMs
11 reports on the NVIDIA RTX 5090 Laptop 24GB and 348 on the AMD Strix Halo 128GB
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the NVIDIA RTX 5090 Laptop 24GB has 33.5 generation tokens/s (2 runs); the AMD Strix Halo 128GB has 13.1 generation tokens/s (4 runs); prompt processing 378.2 tokens/s (3 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 3.5x higher on the NVIDIA RTX 5090 Laptop 24GB (896 GB/s on the NVIDIA RTX 5090 Laptop 24GB, 256 on the AMD Strix Halo 128GB). Across 2 shared models with generation measurements on both cards, the NVIDIA RTX 5090 Laptop 24GB's median is higher on 2 and the AMD Strix Halo 128GB's on 0.
- Memory
- 24GB
- Memory bandwidth
- 896 GB/s
- FP16 compute
- 200 TFLOPS
- Reports
- 11
- Memory
- 128GB unified
- Memory bandwidth
- 256 GB/s
- FP16 compute
- 59 TFLOPS
- Reports
- 348
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 | NVIDIA RTX 5090 Laptop 24GB | AMD Strix Halo 128GB |
|---|---|---|
| Qwen3.8 27B4-bit | 33.52 runsllama.cpp, 1 not stated64K to 100K context | 13.14 runsPP 378.2 (3 runs)llama.cpp, 1 not stated32K context, 3 not stated |
| Qwen3-Coder 30B · 3B active4-bit | 107.0one runPP 3706.5 (one run)llama.cpp64K context | 69.96 runsPP 401.5 (5 runs)llama.cpp8K context, 5 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 | NVIDIA RTX 5090 Laptop 24GB | AMD Strix Halo 128GB |
|---|---|---|
| Qwen3.8 27B | 32.4 t/sQ4_K_M | 3.5 t/sFP16 |
| Qwen3-Coder 30B · 3B active | 128.3 t/sQ4_K_S | 12.6 t/sFP16 |
| Qwen3.8 125B · 6B active | 19.4 t/sQ2_K, experts in RAM | 12.6 t/sQ6_K |
| Qwen3.6 35B · 3B active | 115.0 t/sQ4_K_S | 12.6 t/sFP16 |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | 33.5 t/sQ2_K |
| DeepSeek V4.1 Flash 552B · 16B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 27B | 32.4 t/sQ4_K_M | 3.5 t/sFP16 |
| Gemma 4 26B · 4B active | 73.2 t/sQ5_K_S | 9.5 t/sFP16 |
Change the context, RAM or model in the calculator: NVIDIA RTX 5090 Laptop 24GB · AMD Strix Halo 128GB
Every report on each card, with its source: NVIDIA RTX 5090 Laptop 24GB · AMD Strix Halo 128GB