Explore this comparison’s ranking

Comparison Counted

Which engine runs LFM2.5 1.2B Instruct best on M5 Max · 64 GiB?

Device
M5 Max · 64 GiB (Apple silicon Macs)
OS
macOS 27
Model
LFM2.5 1.2B Instruct, 1.2B parameters
Weights
bf16
Workload
Archive 256 input / 128 output: 256 tokens in, 128 out
Decoding
Greedy, batch 1, context 4,096
Power
Plugged in

Balanced: Decode speed, time to first token and peak memory, weighted equally.

Which engine runs LFM2.5 1.2B Instruct best on M5 Max · 64 GiB? Ranked by Balanced.
#EngineBalanced score out of 100
1MLX LM, an approximate comparison, see details Apple (ml-explore) 93 /100 Balanced score
2Ollama, an approximate comparison, see details Ollama 90 /100 Balanced score
3Magnitude, an approximate comparison, see details Magnitude 89 /100 Balanced score
4Rapid-MLX, an approximate comparison, see details Rapid-MLX contributors 88 /100 Balanced score
5llama.cpp, an approximate comparison, see details ggml-org 87 /100 Balanced score
6ZML, an approximate comparison, see details ZML 78 /100 Balanced score
details

Balanced Decode speed, time to first token and peak memory, weighted equally.

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, magnitude default, model-dtype, zml default.

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

† Magnitude: approximate comparison. The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.

† Rapid-MLX: approximate comparison. The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default.

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

† ZML: approximate comparison. The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.

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 · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default.199.0 range197.4–202.230.7 range30.4–30.72.44 range2.42–2.458,473 range8,453–8,5306of 6 planned
Ollama 0.40.13 · Metal · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default.199.6 range199.4–200.333.8 range33.0–35.12.43 range2.43–2.457,900 range7,605–8,0856of 6 planned
Magnitude 0.2.17 · Metal · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.199.6 range194.8–200.428.7 range19.6–31.02.95 range2.94–2.968,926 range8,265–13,0766of 6 planned
Rapid-MLX 1.0.2 · Metal · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default.198.0 range194.6–199.032.7 range31.1–34.02.70 range2.67–2.707,969 range7,656–8,3876of 6 planned
llama.cpp b12110 · Metal · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default.190.5 range177.9–194.036.7 range35.8–38.52.43 range2.43–2.457,122 range6,779–7,2976of 6 planned
ZML 20261106.0 · Metal · BF16 Eligible, approximately: The experiments use different effective KV cache precision: f16, magnitude default, model-dtype, zml default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.192.6 range147.6–193.324.9 range24.8–26.54.93 range4.72–4.9910,269 range9,659–10,3366of 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.

Decode speed tok/s, higher is better
Time to first token ms, lower is better
Peak memory GiB, lower is better
Prefill speed tok/s, higher is better

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--lfm2.5-1.2b-instruct--bf16--archive-256-128.