Explore this comparison’s ranking

Comparison Counted

Which engine runs Qwen3.8 27B best on M5 Max · 64 GiB?

Device
M5 Max · 64 GiB (Apple silicon Macs)
OS
macOS 27
Model
Qwen3.8 27B, 27B parameters
Weights
4-bit
Workload
Archive 256 input / 128 output: 256 tokens in, 128 out
Decoding
Greedy, batch 1, context 4,096
Power
Plugged in

Memory: Peak memory while generating: the less, the better.

Which engine runs Qwen3.8 27B best on M5 Max · 64 GiB? Ranked by Memory.
#EngineMemory and Decode
1MLX LM, an approximate comparison, see details Apple (ml-explore) Memory: 15.48 GiB 33.4 tok/s decode
2Ollama, an approximate comparison, see details Ollama Memory: 18.90 GiB 26.0 tok/s decode
3llama.cpp, an approximate comparison, see details ggml-org Memory: 18.90 GiB 25.1 tok/s decode
Not ranked in this view
Not rankedMagnitude Magnitude Unsupported Unsupported in 1 comparison. Magnitude 0.2.17 reports this GGUF as Unsupported (magnitude catalog inspect)
Not rankeduzu Mirai Unsupported Unsupported in 1 comparison. 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).
details

Memory Peak memory while generating: the less, the better.

Score: 100 is the best result in every comparison an engine entered; 80 is 80% of the best, on average. Only complete experiments with a passed output check count. How rankings work

† MLX LM: approximate comparison. The experiments use different effective KV cache precision: f16, model-dtype. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.

† Ollama: approximate comparison. The experiments use different effective KV cache precision: f16, model-dtype.

† llama.cpp: approximate comparison. The experiments use different effective KV cache precision: f16, model-dtype.

Every experiment

Medians over valid trials, with the lowest and highest beneath. Eligible experiments are complete and passed their output check. The ranking above selects each engine's configuration for this objective when its required metrics are measured.

Every experiment in this comparison
ExperimentRanking eligibilityDecode tok/s, higher is betterFirst token ms, lower is betterMemory GiB, lower is betterPrefill tok/s, higher is betterValid trials
MLX LM 0.32.2 · Metal · mlx-affine-4bit-g64 Eligible, approximately: The experiments use different effective KV cache precision: f16, model-dtype. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.33.4 range33.4–33.5350 range349–35015.48 range15.47–15.48734.1 range733.2–735.56of 6 planned
Ollama 0.40.13 · Metal · Q4_K_M Eligible, approximately: The experiments use different effective KV cache precision: f16, model-dtype.26.0 range25.7–26.0394 range393–42318.90 range18.89–19.08652.0 range608.6–654.46of 6 planned
llama.cpp b12110 · Metal · Q4_K_M Eligible, approximately: The experiments use different effective KV cache precision: f16, model-dtype.25.1 range23.9–25.3421 range408–43718.90 range18.90–19.08609.7 range586.3–628.86of 6 planned
Magnitude 0.2.17 · Metal · Q4_K_M Unsupported Magnitude 0.2.17 reports this GGUF as Unsupported (magnitude catalog inspect)––––0of 6 planned
uzu 0.6.17 · Metal · mirai-m Unsupported 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).––––0of 6 planned

Every trial, metric by metric

The line runs from an experiment's lowest to its highest valid trial, the dot is the median and each tick is one trial. Every scale starts at zero.

Caveats

  • Publisher cards identify the base model but do not always pin the exact export source checkpoint revision.
  • Engines may differ in quantization scheme, effective KV cache precision, kernels, and activations.

About this comparison

Worked out from its experiments: every experiment here shares the device, the OS major version, the power state, the base model, the quantization class and the workload. Measured under methodology v0.6. Comparison id glenn-m5-max-64gb--macos-27--qwen3.8-27b--4-bit--archive-256-128.