M1 Ultra 128GB vs NVIDIA RTX 4070 for local LLMs
4 reports on the M1 Ultra 128GB and 9 on the NVIDIA RTX 4070
Which is faster for local LLMs?
On Gemma 4 8B at 4-bit, the M1 Ultra 128GB has 69.7 generation tokens/s (one run); prompt processing 1172.9 tokens/s (one run); the NVIDIA RTX 4070 has 55.0 generation 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 1.6x higher on the M1 Ultra 128GB (819 GB/s on the M1 Ultra 128GB, 504 on the NVIDIA RTX 4070).
- Memory
- 128GB unified
- Memory bandwidth
- 819 GB/s
- FP16 compute
- 21 TFLOPS
- Reports
- 4
- Memory
- 12GB
- Memory bandwidth
- 504 GB/s
- FP16 compute
- 116 TFLOPS
- Reports
- 9
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 | M1 Ultra 128GB | NVIDIA RTX 4070 |
|---|---|---|
| Gemma 4 8B4-bit | 69.7one runPP 1172.9 (one run)llama.cppcontext not stated | 55.0one runOllamacontext 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 | M1 Ultra 128GB | NVIDIA RTX 4070 |
|---|---|---|
| Gemma 4 8B | 25.2 t/sFP16 | 42.6 t/sQ6_K |
| Qwen3.8 27B | 7.5 t/sFP16 | 1.7 t/sQ8_0, layers on CPU |
| Qwen3.8 125B · 6B active | 34.0 t/sQ5_K_S | doesn't fit at these settings |
| DeepSeek V4 Flash 284B · 13B active | 53.5 t/sQ2_K | doesn't fit at these settings |
| Qwen3.6 35B · 3B active | 23.5 t/sFP16 | 13.9 t/sQ6_K, experts in RAM |
| Qwen3.6 27B | 7.5 t/sFP16 | 1.7 t/sQ8_0, layers on CPU |
| Gemma 4 26B · 4B active | 20.2 t/sFP16 | 82.3 t/sQ2_K |
Change the context, RAM or model in the calculator: M1 Ultra 128GB · NVIDIA RTX 4070
Every report on each card, with its source: M1 Ultra 128GB · NVIDIA RTX 4070