Rankings

Explore benchmarks

Filter by model, hardware, workload and quantization. Every score links to the comparisons and experiments behind it.

Clear filters

Top Engines

Top Engines, Mobile, Speed score.
#EngineSpeed score out of 100
1ExecuTorch, an approximate comparison, see details PyTorch (Meta) 100 /100 Score
2llama.cpp, an approximate comparison, see details ggml-org 99 /100 Score
Not ranked in this view
Not rankedLiteRT-LM Google AI Edge No rival yet
Not rankeduzu Mirai Failed
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

† ExecuTorch: approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

† llama.cpp: approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

LiteRT-LM: No other engine has a result in the same comparison yet.

uzu: Failed in 1 comparison. Partial: 4 of 6 planned trials were valid.

Coverage

Speed score per device, out of 100: each device ranked on its own.
DesktopMobile
M5 Max 64 GiB macOS 27 Apple M5 MaxOnePlus 13 16 GB Android 16 Qualcomm Snapdragon 8 Elite (SM8750)Pixel 8 8 GB Android 16 Google Tensor G3iPhone 12 4 GB iOS 27 Apple A14 Bionic
llama.cpp76 out of 100, in 3 of 3 comparisons99 out of 100, in 1 of 1 comparisonsFailed Failed in 1 comparison. Partial: 4 of 6 planned trials were valid.No rival yet No other engine has a result in the same comparison yet.
ExecuTorchNot ranked. Results in 1 of 3 comparisons; a rank needs 2.100 out of 100, in 1 of 1 comparisonsNot counted Partial: 2 of 6 planned trials were valid.Not counted Partial: 4 of 6 planned trials were valid.
Magnitude89 out of 100, in 2 of 3 comparisonsNot measured on this device.Not measured on this device.Not measured on this device.
MLX LM97 out of 100, in 2 of 3 comparisonsNot measured on this device.Not measured on this device.Not measured on this device.
Ollama79 out of 100, in 3 of 3 comparisonsNot measured on this device.Not measured on this device.Not measured on this device.
LiteRT-LMNo rival yet No other engine has a result in the same comparison yet.No rival yet No other engine has a result in the same comparison yet.Not counted Its output check failed: 9/12 of the checks the model passes at BF16; needs 11 and the mandatory ones (needle-start, needle-end, continuation-128). Left out, failed by every BF16 configuration: none.Not measured on this device.
uzuNot ranked. Results in 1 of 3 comparisons; a rank needs 2.Not measured on this device.Not measured on this device.Failed Failed in 1 comparison. Partial: 4 of 6 planned trials were valid.
Rapid-MLXNot ranked. Results in 1 of 3 comparisons; a rank needs 2.Not measured on this device.Not measured on this device.Not measured on this device.

Speed

Speed, LFM2.5 1.2B Instruct on OnePlus 13 · 16 GB, 4-bit · Archive 2048 input / 128 output · Android 16
RankEngineDecode and TTFT
1ExecuTorch, approximate comparison, see details PyTorch (Meta)52.1 tok/s TTFT not measured
2llama.cpp, approximate comparison, see details ggml-org51.5 tok/s 8,117 ms TTFT
Not rankedLiteRT-LM Google AI EdgeNot counted
details

LFM2.5 1.2B Instruct on OnePlus 13 · 16 GB · 4-bit · Archive 2048 input / 128 output · Android 16

Medians with the model warmed; model loading is excluded. Each table uses one matched comparison.

ExecuTorch: † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

llama.cpp: † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

LiteRT-LM: Its output check failed: 9/12 of the checks the model passes at BF16; needs 11 and the mandatory ones (needle-start, needle-end, continuation-128). Left out, failed by every BF16 configuration: none. † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

Memory usage by engine

Memory usage by engine, LFM2.5 1.2B Instruct on OnePlus 13 · 16 GB, 4-bit · Archive 2048 input / 128 output · Android 16
RankEngineMemory
Not rankedExecuTorch PyTorch (Meta)Not measured
Not rankedLiteRT-LM Google AI EdgeNot counted
Not rankedllama.cpp ggml-orgNot measured
details

LFM2.5 1.2B Instruct on OnePlus 13 · 16 GB · 4-bit · Archive 2048 input / 128 output · Android 16

Memory ranks require comparable, calibrated measurements. Unavailable measurements remain unranked.

ExecuTorch: No peak memory measured. † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

LiteRT-LM: Its output check failed: 9/12 of the checks the model passes at BF16; needs 11 and the mandatory ones (needle-start, needle-end, continuation-128). Left out, failed by every BF16 configuration: none. † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

llama.cpp: No peak memory measured. † Approximate comparison. The experiments use different effective KV cache precision: fp32, litert-lm default, f16. The experiments here use different 4-bit schemes: 8da4w, litert-int4, Q4_K_M.

Every comparison

Comparisons in this view: each engine's score in each comparison, out of 100, or why it has none.
ComparisonExecuTorchllama.cppLiteRT-LMuzu
LFM2.5 1.2B Instruct on OnePlus 13 · 16 GB 4-bit · Archive 2048 input / 128 output · Android 16 100, approximate99, approximateNot countedNot run
LFM2.5 1.2B Instruct on iPhone 12 · 4 GB 4-bit · Archive 2048 input / 128 output · iOS 27 Not counted: fewer than two scored enginesNot countedMeasured, approximateNot runNot counted
LFM2.5 2.6B on iPhone 12 · 4 GB 4-bit · Archive 2048 input / 128 output · iOS 27 Not counted: fewer than two scored enginesNot runNot countedNot runUnsupported
LFM2.5 350M on iPhone 12 · 4 GB 4-bit · Archive 2048 input / 128 output · iOS 27 Not counted: fewer than two scored enginesNot countedMeasured, approximateNot runFailed
LFM2.5 2.6B on OnePlus 13 · 16 GB 4-bit · Archive 2048 input / 128 output · Android 16 Not counted: fewer than two scored enginesNot runNot countedMeasured, approximateNot run
LFM2.5 350M on OnePlus 13 · 16 GB 4-bit · Archive 2048 input / 128 output · Android 16 Not counted: fewer than two scored enginesNot countedMeasured, approximateNot countedNot run
LFM2.5 1.2B Instruct on Pixel 8 · 8 GB 4-bit · Archive 2048 input / 128 output · Android 16 Not counted: fewer than two scored enginesNot countedNot countedNot countedNot run
LFM2.5 2.6B on Pixel 8 · 8 GB 4-bit · Archive 2048 input / 128 output · Android 16 Not counted: fewer than two scored enginesNot runFailedNot countedNot run
LFM2.5 350M on Pixel 8 · 8 GB 4-bit · Archive 2048 input / 128 output · Android 16 Not counted: fewer than two scored enginesNot countedNot countedNot countedNot run

Each engine’s Speed score in each matched comparison, out of 100, or why it has none. How rankings work