Experiment Partial Published
llama.cpp b12110 on M5 Max · 64 GiB, Metal
LFM2.5 1.2B Instruct as Q4_K_M (4-bit), 2048 input / 128 output: a 2048-token prompt and 128 tokens out. Measured 7 Nov 2026, 18:38 UTC by Glenn Sonna, Xybrid.
Doesn’t count in rankings: Partial: 3 of 6 planned trials were valid.
Experiment id liquid-mac-session1-gguf-q4-2048
Errors
- trial, trial 4: Formal trial requires AC power, low power mode off, nominal thermal state.
- trial, trial 5: Formal trial requires AC power, low power mode off, nominal thermal state.
- trial, trial 6: Formal trial requires AC power, low power mode off, nominal thermal state.
Summary
Over 3 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 | 421.8 | 417.0–436.0 | 425.0 ± 9.9 | 419.4–428.9 |
| Time to first token ms, lower is better | 137 | 137–138 | 138 ± 0.6 | 137–138 |
| Peak memory GiB, lower is better | Not measured | |||
| Prefill speed tok/s, higher is better | 15,026 | 14,937–15,064 | 15,009 ± 65.4 | 14,981–15,045 |
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 | 137 | 15,0642,048 tok in 136 ms, engine | 417.0127 tok in 305 ms, engine | not recorded | nominal → nominal | |
| 2 | 137 | 15,0262,048 tok in 136 ms, engine | 421.8127 tok in 301 ms, engine | not recorded | nominal → nominal | |
| 3 | 138 | 14,9372,048 tok in 137 ms, engine | 436.0127 tok in 291 ms, engine | not recorded | nominal → nominal | |
| 4Invalid | not recorded | not recorded | not recorded | not recorded | The engine returned an error. | |
| 5Invalid | not recorded | not recorded | not recorded | not recorded | The engine returned an error. | |
| 6Invalid | not recorded | not recorded | not recorded | not recorded | The engine returned an error. |
Caveats
- Publisher cards identify the base model but do not pin the exact export source checkpoint revision; exact source-weight identity is unverified.
- GGUF and MLX may differ in quantization, 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
- llama.cpp
- Maintainer
- ggml-org
- License
- MIT, open source
- Version
- b12110
- Commit
- cd959bae9
- Binary SHA-256
- e7e3c60a5e612460d095bfd219047040d6bb9d0b56466aae30a480980b676959
- Built from
- release
- Build flags
- None
- Backend
- Metal
- Threads
- 18
- Interface
- http-server
- Settings in effect
- context_length=4096 parallel=1 speculative=false backend=metal launch.log_verbosity=4 launch.log_colors=off launch.cache_prompt=false launch.cache_ram_mib=0 launch.cache_reuse=0 launch.continuous_batching=false launch.context_shift=false launch.flash_attn=on launch.cache_type_k=f16 launch.cache_type_v=f16 launch.gpu_layers=999 launch.load_mode=mmap offloaded_layers=17 launch.mmap=true n_batch=2048 n_ubatch=512 n_threads=18 n_threads_batch=18 flash_attn=enabled
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-GGUF@8ed288026e23958ad9dfa92d53ed773a8eee7125
- Format
- gguf
- File 1
- LFM2.5-1.2B-Instruct-Q4_K_M.gguf · 697.0 MiB · sha256 b1b3de114215d9507409a662a501a631095a479a419584e8a2ded6304b19b4f5
- File 2
- README.md · 3,544 bytes · sha256 f89d69dd107c19936b98ff981456302bc59255cce34b216ecc489d7bafa782e9
- Quantization
- Q4_K_M (4-bit)
- Weight bits
- 4
- Group size
- Not recorded
- KV cache
- f16
- Conversion
- Published upstream by LiquidAI
- Conversion command
- Not recorded
Workload
- Workload
- 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
- 7 Nov 2026, 18:38 UTC
- Harness
- xybrid-bench 0.3.0 (91cf61a477a49098f1f1352b85e176260af1d1c9)
- Methodology
- v0.3
- Publication
- published
- Published
- 16 Nov 2026, 10:23 UTC
- Reviewed by
- Glenn Sonna
Output check
- Result
- Passed
- Method
- sanity-v1
- 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 19dd83f117525b931fccdafe808aaa6939af792ee549f359eff13fac0d622f5d
- 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; needs 11 and all mandatory checks.
Reproduce it
pnpm bench run --plan plans/liquid-mac-session1.json
Three ABBA blocks; load, two warmups, one trial, unload per turn. Memory remains diagnostic until calibrated.
Raw data
- This record, as committed JSON, schema v0.3.0
- Complete finalized session evidence raw ·
sha256 ba4e27017452…