Llama 3.2 1B Instruct › Apple M3 Max (40-core GPU)
Llama 3.2 1B Instruct on the Apple M3 Max (40-core GPU)
Best measured: 146 tokens/s generation and 3,292 tokens/s prompt processing (Q4_K_M, llamafile).
For comparison, our size-scaling estimate for this combination is 282 tok/s.
Every run
| Quant | Backend | Generate tok/s | Prompt tok/s | Test | Build / by / notes | Source |
|---|---|---|---|---|---|---|
| Q4_K_M | llamafile | 150 | 3,382 | LocalScore avg | LocalScore rank 108/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 146 | 3,292 | LocalScore avg | LocalScore rank 110/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 121 | 2,751 | LocalScore avg | LocalScore rank 131/878 | LocalScore (localscore.ai) |
tg128 = generating 128 tokens; pp512 = reading a 512-token prompt (llama-bench). LocalScore averages several prompt and output lengths, so its numbers run lower than tg128.
No card yet?
Try it on a rented GPU (RunPod)Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.
All results for the Apple M3 Max (40-core GPU) · Llama 3.2 1B Instruct on other hardware