LFM2.5 1.2B Instruct on M5 Max · 64 GiB

Experiment Complete Published

Rapid-MLX 1.0.2 on M5 Max · 64 GiB, Metal

LFM2.5 1.2B Instruct as mlx-affine-8bit-g64 (8-bit), archive 256 input / 128 output: a 256-token prompt and 128 tokens out. Measured 12 Nov 2026, 21:14 UTC by Glenn Sonna, Xybrid.

Eligible for ranking when the objective's required metrics are measured and this configuration is selected.Approximate comparison: The experiments use different effective KV cache precision: litert-lm default, f16, magnitude default, model-dtype, uzu default. Its weights are 8-bit in another scheme: mlx-affine-8bit-g64, where most use Q8_0.

Experiment id s3-lfm2.5-1.2b-8-bit-rapid-mlx-256

Summary

Over 6 valid trials of 6 attempted (6 planned), computed from the raw trials below.

MetricMedianRangeMean ± SDQuartiles
Decode speed tok/s, higher is better361.9 358.8–364.8 361.9 ± 1.9 361.5–362.4
Time to first token ms, lower is better33.5 31.3–37.3 33.8 ± 2.0 33.1–34.0
Peak memory GiB, lower is better1.69 1.68–1.70 1.69 ± 0.01 1.68–1.70
Prefill speed tok/s, higher is better7,775 6,968–8,328 7,730 ± 443.0 7,666–7,873

Trials, metric by metric

The line runs from the lowest to the highest valid trial, the dot is the median and each tick is one trial.

Every trial

Raw token counts and durations, with the speeds computed from them. Invalid trials stay in the record with their reason; they are left out of the summary, never silently dropped.

Every attempted trial
TrialTime to first token msPrefill tok/sDecode tok/sPeak memory GiBThermalNote
137.36,968256 tok in 36.7 ms, engine358.8127 tok in 354 ms, engine1.69nominal → nominal
234.17,653256 tok in 33.5 ms, engine362.4127 tok in 350 ms, engine1.70nominal → nominal
331.38,328256 tok in 30.7 ms, engine361.6127 tok in 351 ms, engine1.69nominal → nominal
433.27,844256 tok in 32.6 ms, engine361.5127 tok in 351 ms, engine1.70nominal → nominal
533.87,706256 tok in 33.2 ms, engine362.3127 tok in 351 ms, engine1.68nominal → nominal
633.17,883256 tok in 32.5 ms, engine364.8127 tok in 348 ms, engine1.68nominal → nominal

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.

Configuration

Device

Device
Apple Mac17,6
Cohort
Apple silicon Macs (desktop)
SoC
Apple M5 Max
CPU
18-core Apple CPU
GPU
Not recorded
NPU
Not recorded
RAM
64.00 GiB
OS
macOS 27.2.0 (26B123)
Kernel
27.2.0
Power
mains
Low power mode
No
Performance mode
Not recorded

Engine

Engine
Rapid-MLX
Maintainer
Rapid-MLX contributors
License
Apache-2.0, open source
Version
1.0.2
Commit
f1aa011aa91bb38140ed3fb68f6267b77447ee0f
Binary SHA-256
d5c9b638732cda0b2e2a816aa85cc1780e0a3c1b690bb23c8d9b4d3ac81a527c
Built from
package
Build flags
None
Backend
Metal
Threads
Not recorded
Interface
python-worker
Settings in effect
backend=Device(gpu, 0) model_source=artifact folder context_length=4096 batch_size=1 max_num_seqs=1 max_concurrent_requests=1 prefill_batch_size=1 completion_batch_size=1 prefill_step_size=2048 prefix_cache=false disk_caches=false speculative=false spec_decode=none suffix_decoding=false response_cache_entries=0 pflash=off kv_cache_dtype=bf16 kv_quantized=false admission_window_ms=0 stream_interval=1 gpu_memory_utilization=0.9 greedy=true temperature=0 sampler=argmax trust_remote_code=false telemetry=false weight_bits=8 weight_group_size=64

Model and weights

Base model
LFM2.5 1.2B Instruct (LFM2.5, 1.2B)
Upstream
LiquidAI/LFM2.5-1.2B-Instruct@b9d816ab510f9a28dca37cb182e9da8b0226e13c
Weights
LiquidAI/LFM2.5-1.2B-Instruct-MLX-8bit@ce5150fd0de58d2bba8d88cc87f13e4fbda5977a
Format
safetensors
File 1
README.md · 2,286 bytes · sha256 e121410071d946a17d3481d16ddf5548a335433be9a30eafe1532265ddf84fbd
File 2
chat_template.jinja · 1,836 bytes · sha256 3acb41f339d069a43037e6b9a9715cbeece8874eb25a7cabf1a33903d7724d6d
File 3
config.json · 1,572 bytes · sha256 07f8d0fa634323c35c91978f38fa7ffe9bc50439a92a57dadcd81e0e69aaa501
File 4
generation_config.json · 132 bytes · sha256 5ffd97da1dec4308543894569662d96e923ed01f7a9d8c7ff5aea7f800738cbd
File 5
model.safetensors · 1.16 GiB · sha256 74e6e29f0215107a4aa9d4bd0a1ac8c1adbd0096f898bfbf49d3edc6cc7e4130
File 6
model.safetensors.index.json · 23,415 bytes · sha256 fcfe171be23642632ad0e997d641bb2a92c137635660e28f07efa4eb3dbe659b
File 7
special_tokens_map.json · 434 bytes · sha256 742aefe2b7dec496e8caffdba03a75d0c1a9925d53bd3f3e0d388c96b591b6f4
File 8
tokenizer.json · 4.5 MiB · sha256 df1d8d5ec5d091b460562ffd545e4a5e91d17d4a0db7ebe733be34ed374377bd
File 9
tokenizer_config.json · 92,225 bytes · sha256 2a52ec012d3df831ba434b081bef3726a6ee22501f062ad8353c557a0cfa0d01
Quantization
mlx-affine-8bit-g64 (8-bit)
Weight bits
8
Group size
64
KV cache
model-dtype
Conversion
Published upstream by LiquidAI
Conversion command
Not recorded

Workload

Workload
Archive 256 input / 128 output
Prompt
256 tokens, from prompts/source.txt
Prompt SHA-256
c10d1b7bef62d95322f8f8d229dc66118d9375107bdc49b06f76287d058b2f6c
Chat template
ba551d58630afa3190b1be3602e28301f3d2e9bbac978dfc49d6d825171648b6
Output
128 tokens, ignoring end of sequence
Context length
4,096
Batch size
1
Sampling
Greedy
Timing
One streamed request per trial
Warmup
2 runs, not counted
Trials planned
6
Cooldown
15 s between trials

Conditions

Ambient
Not recorded
Thermal at start
nominal
Thermal at end
nominal
Thermal gate
macOS thermal state nominal
Throttling detected
No

Provenance

Measured by
Glenn Sonna, Xybrid
Operator affiliation
Xybrid
Run at
12 Nov 2026, 21:14 UTC
Harness
xybrid-bench 0.4.0 (18f43990b479e8cd3ab97664a2cca0d707214164)
Methodology
v0.6
Publication
published
Published
16 Nov 2026, 10:23 UTC
Reviewed by
Glenn Sonna

Output check

Result
Passed
Method
sanity-v1-bf16-reference
Reference
checks/sanity-v1.json
Reference SHA-256
a702dee079525ae0ec90465b1c9e26052c99cb3e1c53e83dea05be4c194fc7ca
Required passing checks
11
Check arithmetic
Passed · sha256 6b51d431df5d7f141cbececcf79edf3dd861c3b4069f0b11661a3eefacbba918
Check capital
Passed · sha256 5dd272b4f316b776a7b8e3d0894b37e1e42be3d5d3b204b8a5836cc50597a6b1
Check opposite
Passed · sha256 f151cd9e12c8994edfa316e4624f956bc68460454869177f287146f8cc956a12
Check weekday
Passed · sha256 7d8af1de1262f150187d5938bf649d6c35726970685f916d1ec6392a801b9762
Check color
Passed · sha256 d486dfbd5fb578340bccbdd0a662527eab38550648d5f44517e5ac71b8824703
Check sort
Passed · sha256 a386b3c9b4b4786df5bb6474bab0a62b8476e2f3f9c8a6433aca40152840f6b7
Check animal
Passed · sha256 4b227777d4dd1fc61c6f884f48641d02b4d121d3fd328cb08b5531fcacdabf8a
Check word
Passed · sha256 3b635d3b4437e023e72a9df26201b6ab51285b376446446ebb379121994550f6
Check copy
Passed · sha256 d61e6890d0e3571b0c7693612d41f8d059a1fcdef4ba959c0af4144541d41907
Check needle-start
Passed · sha256 8b9af2e5b8b4cfddb87d171997b95c471f8d68a68e30fb5177abeb3332372edd
Check needle-end
Passed · sha256 e6589c736af32f81bf239c5eda6de41959122ccc1d05463b75e0eb0c70fc831c
Check continuation-128
Passed · sha256 fe1fdf5d49caa68c1ef2f8a341413cb2c3826023aef1113c6c289dcd86193b32
Detail
12/12 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.

Reproduce it

Commands
pnpm bench run --plan plans/s3-lfm2.5-1.2b-8-bit.json

Three palindromic blocks per quantization class and workload; load, two warmups, one trial, unload per turn.

Raw data