Llama 3.2 1B Instruct › AMD Radeon 780M
Llama 3.2 1B Instruct on the AMD Radeon 780M
Best measured: 11.0 tokens/s generation and 582 tokens/s prompt processing (Q4_K_M, llamafile).
For comparison, our size-scaling estimate for this combination is 18.8 tok/s.
Every run
| Quant | Backend | Generate tok/s | Prompt tok/s | Test | Build / by / notes | Source |
|---|---|---|---|---|---|---|
| Q4_K_M | llamafile | 11.7 | 690 | LocalScore avg | LocalScore rank 367/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 12.1 | 632 | LocalScore avg | LocalScore rank 382/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 11.0 | 640 | LocalScore avg | LocalScore rank 390/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 11.5 | 617 | LocalScore avg | LocalScore rank 395/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 11.2 | 572 | LocalScore avg | LocalScore rank 418/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 10.3 | 592 | LocalScore avg | LocalScore rank 419/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 11.0 | 567 | LocalScore avg | LocalScore rank 427/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 7.5 | 506 | LocalScore avg | LocalScore rank 496/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 9.5 | 334 | LocalScore avg | LocalScore rank 575/878 | LocalScore (localscore.ai) |
| Q4_K_M | llamafile | 5.9 | 399 | LocalScore avg | LocalScore rank 581/878 | 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 AMD Radeon 780M · Llama 3.2 1B Instruct on other hardware