Rankings
Explore benchmarks
Filter by model, hardware, workload and quantization. Every score links to the comparisons and experiments behind it.
Top Engines
No comparisons match these filters.
Coverage
| Desktop | |
|---|---|
| M5 Max 64 GiB macOS 27 Apple M5 Max | |
| llama.cpp | 864/4 out of 100, in 4 of 4 comparisons |
| MLX LM | 984/4 out of 100, in 4 of 4 comparisons |
| Ollama | 884/4 out of 100, in 4 of 4 comparisons |
| Magnitude | Unsupported Unsupported in 4 comparisons. Magnitude 0.2.17 reports this GGUF as Unsupported (magnitude catalog inspect) |
| uzu | Unsupported Unsupported in 4 comparisons. uzu 0.6.17 renders its own Qwen3.8 chat template: 67 input tokens for the 27-token canonical check prompt, so it can't run the frozen prompts (measured in qualification). |
Speed
Not measured.
Memory usage by engine
Not measured.