NVIDIA L4 vs M3 Ultra 512GB for local LLMs
3 reports on the NVIDIA L4 and 41 on the M3 Ultra 512GB
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the NVIDIA L4 has 14.6 generation tokens/s (one run); the M3 Ultra 512GB has 31.8 generation tokens/s (one run); prompt processing 391.2 tokens/s (one run). 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 2.7x higher on the M3 Ultra 512GB (300 GB/s on the NVIDIA L4, 819 on the M3 Ultra 512GB).
- Memory
- 24GB
- Memory bandwidth
- 300 GB/s
- FP16 compute
- 121 TFLOPS
- Reports
- 3
- Memory
- 512GB unified
- Memory bandwidth
- 819 GB/s
- FP16 compute
- 28.4 TFLOPS
- Reports
- 41
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 L4 | M3 Ultra 512GB |
|---|---|---|
| Qwen3.8 27B4-bit | 14.6one runllama.cppcontext not stated | 31.8one runPP 391.2 (one run)MLXcontext 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 L4 | M3 Ultra 512GB |
|---|---|---|
| Qwen3.8 27B | 10.9 t/sQ4_K_M | 8.0 t/sFP16 |
| Qwen3.8 125B · 6B active | 16.4 t/sQ2_K, experts in RAM | 13.6 t/sFP16 |
| Qwen3.6 35B · 3B active | 38.5 t/sQ4_K_S | 28.7 t/sFP16 |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | 22.6 t/sQ8_0 |
| Qwen3.6 27B | 10.9 t/sQ4_K_M | 8.0 t/sFP16 |
| Gemma 4 26B · 4B active | 24.5 t/sQ5_K_S | 21.5 t/sFP16 |
| DeepSeek V4.1 Flash 552B · 16B active | doesn't fit at these settings | 30.0 t/sQ4_K_M |
Change the context, RAM or model in the calculator: NVIDIA L4 · M3 Ultra 512GB
Every report on each card, with its source: NVIDIA L4 · M3 Ultra 512GB