Explore benchmarks
Filter by model, hardware, workload and quantization. Every score links to the comparisons and experiments behind it.
Top Engines
| # | Engine | Speed score out of 100 |
|---|---|---|
| 1 | Rapid-MLX†, an approximate comparison, see details | 100 /100 Score |
| ||
| 2 | MLX LM†, an approximate comparison, see details | 97 /100 Score |
| ||
| 3 | Magnitude†, an approximate comparison, see details | 88 /100 Score |
| ||
| 4 | Ollama†, an approximate comparison, see details | 70 /100 Score |
| ||
| 5 | llama.cpp†, an approximate comparison, see details | 65 /100 Score |
| ||
| Not ranked in this view | ||
| Not ranked | uzu | Too few results |
| Not ranked | ExecuTorch | Not counted |
| Not ranked | LiteRT-LM | Not counted |
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
The score combines counted comparisons. Select one comparison to see its measured values.
† Rapid-MLX: approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
† MLX LM: approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
† Magnitude: approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
† Ollama: approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
† llama.cpp: approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
uzu: Results in 1 of 2 comparisons; a rank needs 2.
ExecuTorch: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128.
LiteRT-LM: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128.
Coverage
| Desktop | Mobile | |||
|---|---|---|---|---|
| M5 Max 64 GiB macOS 27 Apple M5 Max | OnePlus 13 16 GB Android 16 Qualcomm Snapdragon 8 Elite (SM8750) | Pixel 8 8 GB Android 16 Google Tensor G3 | iPhone 12 4 GB iOS 27 Apple A14 Bionic | |
| llama.cpp | 652/2 out of 100, in 2 of 2 comparisons | No 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. | 431/1 out of 100, in 1 of 1 comparisons |
| uzu | 1/2Not ranked. Results in 1 of 2 comparisons; a rank needs 2. | Not measured on this device. | Not measured on this device. | 1001/1 out of 100, in 1 of 1 comparisons |
| Magnitude | 882/2 out of 100, in 2 of 2 comparisons | Not measured on this device. | Not measured on this device. | Not measured on this device. |
| MLX LM | 972/2 out of 100, in 2 of 2 comparisons | Not measured on this device. | Not measured on this device. | Not measured on this device. |
| Ollama | 702/2 out of 100, in 2 of 2 comparisons | Not measured on this device. | Not measured on this device. | Not measured on this device. |
| Rapid-MLX | 1002/2 out of 100, in 2 of 2 comparisons | Not measured on this device. | Not measured on this device. | Not measured on this device. |
| ExecuTorch | Not counted Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not counted Its output check failed: 5/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not counted Its output check failed: 5/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not counted Its output check failed: 5/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. |
| LiteRT-LM | Not counted Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not counted Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not counted Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. | Not measured on this device. |
Speed
| Rank | Engine | Prefill |
|---|---|---|
| 1 | uzu†, approximate comparison, see details | 19,884 tok/s |
| 2 | Magnitude†, approximate comparison, see details | 13,679 tok/s |
| 3 | Ollama†, approximate comparison, see details | 12,795 tok/s |
| 4 | llama.cpp†, approximate comparison, see details | 11,475 tok/s |
| 5 | MLX LM†, approximate comparison, see details | 11,258 tok/s |
| 6 | Rapid-MLX†, approximate comparison, see details | 10,658 tok/s |
| Not ranked | ExecuTorch | Not counted |
| Not ranked | LiteRT-LM | Not counted |
details
LFM2.5 350M on M5 Max · 64 GiB · 4-bit · Archive 256 input / 128 output · macOS 27
Medians with the model warmed; model loading is excluded. Each table uses one matched comparison.
uzu: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mirai-m, where most use Q4_K_M.
Magnitude: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.
Ollama: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
llama.cpp: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
MLX LM: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.
Rapid-MLX: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.
ExecuTorch: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-4w-g64, where most use Q4_K_M.
LiteRT-LM: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: litert-int4, where most use Q4_K_M.
Memory usage by engine
| Rank | Engine | Memory |
|---|---|---|
| 1 | Ollama†, approximate comparison, see details | 0.46 GiB |
| 2 | llama.cpp†, approximate comparison, see details | 0.46 GiB |
| 3 | uzu†, approximate comparison, see details | 0.48 GiB |
| 4 | MLX LM†, approximate comparison, see details | 0.50 GiB |
| 5 | Rapid-MLX†, approximate comparison, see details | 0.72 GiB |
| 6 | Magnitude†, approximate comparison, see details | 1.05 GiB |
| Not ranked | ExecuTorch | Not counted |
| Not ranked | LiteRT-LM | Not counted |
details
LFM2.5 350M on M5 Max · 64 GiB · 4-bit · Archive 256 input / 128 output · macOS 27
Memory ranks require comparable, calibrated measurements. Unavailable measurements remain unranked.
Ollama: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
llama.cpp: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default.
uzu: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mirai-m, where most use Q4_K_M.
MLX LM: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.
Rapid-MLX: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-affine-4bit-g64, where most use Q4_K_M.
Magnitude: † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its phases were timed by the harness clock around the engine API, so they include transport that engine counters leave out.
ExecuTorch: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: mlx-4w-g64, where most use Q4_K_M.
LiteRT-LM: Its output check failed: 6/8 of the checks the model passes at BF16; needs 7 and the mandatory ones (needle-start, needle-end). Left out, failed by every BF16 configuration: arithmetic, sort, word, continuation-128. † Approximate comparison. The experiments use different effective KV cache precision: bf16, fp32, litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 4-bit in another scheme: litert-int4, where most use Q4_K_M.
Every comparison
| Comparison | Rapid-MLX | MLX LM | Magnitude | Ollama | llama.cpp | uzu | ExecuTorch | LiteRT-LM |
|---|---|---|---|---|---|---|---|---|
| LFM2.5 350M on M5 Max · 64 GiB 4-bit · Archive 2048 input / 128 output · macOS 27 | 100†, approximate | 95†, approximate | 93†, approximate | 74†, approximate | 69†, approximate | Failed | Not counted | Not counted |
| LFM2.5 350M on M5 Max · 64 GiB 4-bit · Archive 256 input / 128 output · macOS 27 | 100†, approximate | 100†, approximate | 82†, approximate | 67†, approximate | 62†, approximate | 90†, approximate | Not counted | Not counted |
Each engine’s Speed score in each matched comparison, out of 100, or why it has none. How rankings work