Configs › Qwen3-8B tok/s: Mac Max 110.5, DGX Spark 42
Qwen3-8B tok/s: Mac Max 110.5, DGX Spark 42
Measured Qwen3-8B speeds: 110.5 tok/s on Mac Max (MLX-4BIT, mlx); 42 tok/s on DGX Spark. 3 configs, each linked to its source, snapshot 2026-10-15.
Snapshot 2026-10-15 · 394 configs on AI boxes and 423 on GPU cards · 123 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.
Quick answers
- How fast is Qwen3-8B on Mac Max?
- 110.5 tok/s single-stream decode (MLX-4BIT, mlx), the fastest of 2 measured configs on Apple M-series Max (MacBook Pro / Mac Studio Max), measured by emihiggins, source last checked 2026-09-29. source
- How fast is Qwen3-8B on DGX Spark?
- 42 tok/s single-stream decode (Q4_K_M, llama.cpp), the only measured config on NVIDIA DGX Spark (GB10), measured by DandinPower, source last checked 2026-10-14. source
Measured configs for Qwen3-8B
| Model | Hardware | Quant | Backend | Decode | Context | Trust | Source |
|---|---|---|---|---|---|---|---|
| Qwen3-8B | Mac Max | MLX-4BIT | mlx | 110.5 tok/s (prefill 3151) | 512 | Extracted, source-linked | source |
| Qwen3-8B | Mac Max | Q4_K_M | llama.cpp | 97.7 tok/s (prefill 2745) | 512 | Extracted, source-linked | source |
| Qwen3-8B | DGX Spark | Q4_K_M | llama.cpp | 42 tok/s | 2048 | Extracted, source-linked | source |
More: all configs · hardware · news · contributors · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.