Compare
Loading community benchmarks…
Loading community benchmarks…
Put two chips, runtimes, or runtime versions side by side. Everything else is held fixed or called out, so a ratio is worth exactly what the matched runs say.
Ratios read side A relative to side B. Geometric mean over 2 like-for-like pairs.
Everything else matches: model, quantisation, backend, protocol, and workload are the same on both sides.
Each row is a configuration present on both sides. Values are per-cell medians; ratios read Apple M1 Max relative to Apple M3 Ultra.
| Configuration | Decode A | Decode B | Ratio | Prefill A | Prefill B | Ratio | Runs A / B |
|---|---|---|---|---|---|---|---|
Qwen3.5-2B-Base BaseRTQ4 | 190.0 | 293.8 | 0.65× | 439 | 1,756 | 0.25× | 1 / 1 |
Qwen3.5-2B-Base BaseRTQ8 | 140.7 | 222.4 | 0.63× | 438 | 1,696 | 0.26× | 1 / 1 |
Top 2 of 2 runs by decode throughput.
| Model / format | Chip / backend | Decode | Prefill | Contributor | Date | Report |
|---|---|---|---|---|---|---|
Qwen3.5-2B-Base-Q4 BaseRT 0.2.6Q4 | Apple M1 Max Metal | 190.0 TG128 @ 1 ctx | 439 PP512 | basecompute | View Benchmark | |
Qwen3.5-2B-Base-Q8 BaseRT 0.2.6Q8 | Apple M1 Max Metal | 140.7 TG128 @ 1 ctx | 438 PP512 | basecompute | View Benchmark |
Top 2 of 2 runs by decode throughput.
| Model / format | Chip / backend | Decode | Prefill | Contributor | Date | Report |
|---|---|---|---|---|---|---|
Qwen3.5-2B-Base-Q4 BaseRT 0.2.6Q4 | Apple M3 Ultra Metal | 293.8 TG128 @ 1 ctx | 1,756 PP512 | basecompute | View Benchmark | |
Qwen3.5-2B-Base-Q8 BaseRT 0.2.6Q8 | Apple M3 Ultra Metal | 222.4 TG128 @ 1 ctx | 1,696 PP512 | basecompute | View Benchmark |