Explore this comparison’s ranking

Comparison Counted

Which engine runs LFM2.5 2.6B best on M5 Max · 64 GiB?

Device
M5 Max · 64 GiB (Apple silicon Macs)
OS
macOS 27
Model
LFM2.5 2.6B, 2.6B parameters
Weights
8-bit
Workload
Archive 256 input / 128 output: 256 tokens in, 128 out
Decoding
Greedy, batch 1, context 4,096
Power
Plugged in

Speed: Decode speed: tokens generated per second.

Which engine runs LFM2.5 2.6B best on M5 Max · 64 GiB? Ranked by Speed.
#EngineDecode and First token
1MLX LM, an approximate comparison, see details Apple (ml-explore) Decode: 163.1 tok/s 53.2 ms first token
2Rapid-MLX, an approximate comparison, see details Rapid-MLX contributors Decode: 162.6 tok/s 61.2 ms first token
3Ollama, an approximate comparison, see details Ollama Decode: 155.1 tok/s 52.7 ms first token
4llama.cpp, an approximate comparison, see details ggml-org Decode: 148.8 tok/s 62.8 ms first token
Not ranked in this view
Not rankedMagnitude Magnitude Unsupported Unsupported in 1 comparison. Magnitude's chat API renders LFM2.5 2.6B's own template, which always opens a thinking block, so it streams reasoning and can't run the closed-think prompts (measured in qualification).
Not rankeduzu Mirai Unsupported Unsupported in 1 comparison. uzu can't run the frozen prompts: Error: uzu counted 23 input tokens, not the frozen prompt's 26 (measured in qualification).
Not rankedLiteRT-LM Google AI Edge Not counted Its output check failed: 7/9 of the checks the model passes at BF16; needs 8 and the mandatory ones (needle-start, continuation-128). Left out, failed by every BF16 configuration: color, sort, needle-end.
details

Speed Decode speed: tokens generated per second.

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: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.

† Rapid-MLX: approximate comparison. The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.

† Ollama: approximate comparison. The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.

† llama.cpp: approximate comparison. The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.

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-8bit-g64 Eligible, approximately: The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.163.1 range158.1–163.953.2 range53.0–55.13.02 range2.99–3.074,868 range4,703–4,8826of 6 planned
Rapid-MLX 1.0.2 · Metal · mlx-affine-8bit-g64 Eligible, approximately: The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.162.6 range161.1–163.661.2 range58.9–61.93.28 range3.23–3.414,220 range4,175–4,3896of 6 planned
Ollama 0.40.13 · Metal · Q8_0 Eligible, approximately: The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.155.1 range155.0–155.252.7 range52.1–52.92.96 range2.96–2.974,995 range4,957–5,0516of 6 planned
llama.cpp b12110 · Metal · Q8_0 Eligible, approximately: The experiments use different effective KV cache precision: litert-lm default, f16, model-dtype. The experiments here use different 8-bit schemes: litert-int8, Q8_0, mlx-affine-8bit-g64.148.8 range116.0–150.162.8 range61.2–63.72.96 range2.96–2.984,126 range4,062–4,2416of 6 planned
LiteRT-LM 0.19.1 · CPU · litert-int8 Complete Its output check failed: 7/9 of the checks the model passes at BF16; needs 8 and the mandatory ones (needle-start, continuation-128). Left out, failed by every BF16 configuration: color, sort, needle-end.60.3 range60.0–60.5541 range539–5423.42 range3.42–3.42515.9 range514.4–517.46of 6 planned
Magnitude 0.2.17 · Metal · Q8_0 Unsupported Magnitude's chat API renders LFM2.5 2.6B's own template, which always opens a thinking block, so it streams reasoning and can't run the closed-think prompts (measured in qualification).––––0of 6 planned
uzu 0.6.17 · Metal · mirai-l Unsupported uzu can't run the frozen prompts: Error: uzu counted 23 input tokens, not the frozen prompt's 26 (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.

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--lfm2.5-2.6b--8-bit--archive-256-128.