Experiment Complete Published
ExecuTorch 1.6.0 on M5 Max · 64 GiB, Metal
LFM2.5 1.2B Instruct as mlx-4w-g64 (4-bit), archive 2048 input / 128 output: a 2048-token prompt and 128 tokens out. Measured 12 Nov 2026, 22:07 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: 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.
Experiment id s3-lfm2.5-1.2b-4-bit-et-mlx-2048
Summary
Over 6 valid trials of 6 attempted (6 planned), computed from the raw trials below.
| Metric | Median | Range | Mean ± SD | Quartiles |
|---|---|---|---|---|
| Decode speed tok/s, higher is better | 456.8 | 445.6–458.5 | 454.7 ± 4.7 | 454.4–456.8 |
| Time to first token ms, lower is better | 332 | 330–332 | 331 ± 0.8 | 331–332 |
| Peak memory GiB, lower is better | Not measured | |||
| Prefill speed tok/s, higher is better | 6,178 | 6,169–6,206 | 6,181 ± 15.3 | 6,169–6,187 |
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.
| Trial | Time to first token ms | Prefill tok/s | Decode tok/s | Peak memory GiB | Thermal | Note |
|---|---|---|---|---|---|---|
| 1 | 331 | 6,1872,048 tok in 331 ms, engine | 456.8127 tok in 278 ms, engine | not recorded | nominal → nominal | |
| 2 | 331 | 6,1872,048 tok in 331 ms, engine | 445.6127 tok in 285 ms, engine | not recorded | nominal → nominal | |
| 3 | 332 | 6,1692,048 tok in 332 ms, engine | 456.8127 tok in 278 ms, engine | not recorded | nominal → nominal | |
| 4 | 330 | 6,2062,048 tok in 330 ms, engine | 458.5127 tok in 277 ms, engine | not recorded | nominal → nominal | |
| 5 | 332 | 6,1692,048 tok in 332 ms, engine | 456.8127 tok in 278 ms, engine | not recorded | nominal → nominal | |
| 6 | 332 | 6,1692,048 tok in 332 ms, engine | 453.6127 tok in 280 ms, engine | not recorded | nominal → 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
- ExecuTorch
- Maintainer
- PyTorch (Meta)
- License
- BSD-3-Clause, open source
- Version
- 1.6.0
- Commit
- Not recorded
- Binary SHA-256
- 5ffb7850c43a0ee6d679ba2d94973c6060256a9f05ac14067496498b24ff9095
- Built from
- package
- Build flags
- None
- Backend
- Metal
- Threads
- Not recorded
- Interface
- python-worker
- Settings in effect
- backend=metal recipe=lfm2.5-1.2b-et-mlx-4w,lfm2_mlx_4w.yaml,lfm2.5-1.2b-et-mlx-4w.pte,mlx-4w-g64 delegates=MLXBackend echo=false temperature=0 num_bos=0 num_eos=0 reset_before_generation=true eos_ignorable=false
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@b9d816ab510f9a28dca37cb182e9da8b0226e13c
- Format
- pte
- File 1
- lfm2.5-1.2b-et-mlx-4w.pte · 849.3 MiB · sha256 3a1138df721d3424cfb5beefa076446c3e9d448c8b5d0f8e2444bcbd08c4002b
- File 2
- tokenizer.json · 4.5 MiB · sha256 df1d8d5ec5d091b460562ffd545e4a5e91d17d4a0db7ebe733be34ed374377bd
- Quantization
- mlx-4w-g64 (4-bit)
- Weight bits
- 4
- Group size
- 64
- KV cache
- bf16
- Conversion
- Converted by Xybrid with executorch export_llm 1.6.0
- Conversion command
- cd "$ROOT/engines/executorch/build/lfm2.5-1.2b-et-mlx-4w" && PYTHONPATH="$ROOT/engines/executorch/.venv/lib/python3.12/site-packages/executorch" "$ROOT/engines/executorch/.venv/bin/python" -m executorch.extension.llm.export.export_llm --config "$ROOT/engines/executorch/.venv/lib/python3.12/site-packages/executorch/examples/models/lfm2/config/lfm2_mlx_4w.yaml" +base.model_class=lfm2_5_1_2b "+base.checkpoint=$ROOT/engines/executorch/build/lfm2.5-1.2b-weights.pth" "+base.params=$ROOT/downloads/executorch/4776f71b6f0daa5dfc7d758d5836fc1cc0d40154/examples/models/lfm2/config/lfm2_5_1_2b_config.json" +export.max_seq_length=4096 +export.max_context_length=4096 +export.output_name=lfm2.5-1.2b-et-mlx-4w.pte
Workload
- Workload
- Archive 2048 input / 128 output
- Prompt
- 2,048 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, 22:07 UTC
- Harness
- xybrid-bench 0.4.0 (a101f5a635e0541dbb6a1f784441a4634584f414)
- 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
- Failed · sha256 ec7d56a01607001e6401366417c5e2eb00ffa0df17ca1a9a831e0b32c8f11bf7
- 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
- 11/12 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
pnpm bench run --plan plans/s3-lfm2.5-1.2b-4-bit.json
Three palindromic blocks per quantization class and workload; load, two warmups, one trial, unload per turn.
Raw data
- This record, as committed JSON, schema v0.3.0
- Complete finalized session evidence raw ·
sha256 83ed1948cfa3…