Experiment Complete Published
LiteRT-LM 0.19.0 on Pixel 8 · 8 GB, CPU
LFM2.5 350M as litert-int4 (4-bit), archive 2048 input / 128 output: a 2048-token prompt and 128 tokens out. Measured 12 Nov 2026, 12:23 UTC by Glenn Sonna, Xybrid.
Doesn’t count in rankings: 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.
Experiment id s4-pixel8-lfm2.5-350m-4-bit-litert-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 | 45.4 | 43.0–48.1 | 45.3 ± 2.0 | 43.7–46.7 |
| Time to first token ms, lower is better | Not measured | |||
| Peak memory GiB, lower is better | Not measured | |||
| Prefill speed tok/s, higher is better | 311.9 | 276.2–314.5 | 306.4 ± 14.9 | 310.4–313.5 |
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 | not recorded | 276.22,048 tok in 7,414 ms, engine | 43.4127 tok in 2,926 ms, engine | not recorded | 37.4 °C, nominal → 37.5 °C, nominal | |
| 2 | not recorded | 310.62,048 tok in 6,594 ms, engine | 46.4127 tok in 2,738 ms, engine | not recorded | 35.8 °C, nominal → 37.1 °C, nominal | |
| 3 | not recorded | 314.52,048 tok in 6,512 ms, engine | 46.8127 tok in 2,713 ms, engine | not recorded | 36.4 °C, nominal → 37.2 °C, nominal | |
| 4 | not recorded | 310.32,048 tok in 6,599 ms, engine | 48.1127 tok in 2,643 ms, engine | not recorded | 36.2 °C, nominal → 37.5 °C, nominal | |
| 5 | not recorded | 313.62,048 tok in 6,530 ms, engine | 44.4127 tok in 2,858 ms, engine | not recorded | 37.0 °C, nominal → 38.1 °C, nominal | |
| 6 | not recorded | 313.32,048 tok in 6,538 ms, engine | 43.0127 tok in 2,956 ms, engine | not recorded | 36.6 °C, nominal → 37.8 °C, 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
- Google Pixel 8
- Cohort
- Android phones (mobile)
- SoC
- Google Tensor G3
- CPU
- 9-core CPU (1× Cortex-X3, 4× Cortex-A715, 4× Cortex-A510)
- GPU
- Arm Mali-G715
- NPU
- Not recorded
- RAM
- 8.00 GiB
- OS
- Android 16 (CP1A.260505.005)
- Kernel
- 6.1.145-android14-11-gfa1d6308d1fe-ab14691759
- Power
- mains
- Low power mode
- No
- Wakefulness
- awake
Engine
- Engine
- LiteRT-LM
- Maintainer
- Google AI Edge
- License
- Apache-2.0, open source
- Version
- 0.19.0
- Commit
- Not recorded
- Binary SHA-256
- 8e199c6b3fb438cb94943567a279f9e1f994d57c3e1f4667084fc8ae2d06a6eb
- Built from
- release
- Build flags
- None
- Backend
- CPU
- Threads
- Not recorded
- Interface
- cli
- Settings in effect
- backend=cpu process=one per generation chat_template=none (session API) sampler=top_k=1 temperature=0 launch.max_num_tokens=4096 launch.num_cpu_threads=engine default launch.cache_dir=/data/local/tmp/xybrid/litert-cache launch.benchmark=true launch.benchmark_prefill_tokens=0 launch.benchmark_decode_tokens=0 launch.async=false
Model and weights
- Base model
- LFM2.5 350M (LFM2.5, 350M)
- Upstream
- LiquidAI/LFM2.5-350M@9921cd4d82c0e3986887aa4b376f9446729ef8a3
- Weights
- litert-community/LFM2.5-350M@4754bf697659039afee2962cfd9c8f40d547ab35
- Format
- litertlm
- File 1
- LFM2.5-350M_int4.litertlm · 236.3 MiB · sha256 c29a41cdf4148ec9f4b6e57f8e598728a05a189fdc7ca0764b013b78f9dfd52f
- File 2
- README.md · 14,806 bytes · sha256 4aa8f3ef769b89ecf207b1506da4a8c617cd7aafbcf9520a96ae2e10f1a6bc56
- Quantization
- litert-int4 (4-bit)
- Weight bits
- 4
- Group size
- Not recorded
- KV cache
- litert-lm default
- Conversion
- Published upstream by litert-community
- Conversion command
- Not recorded
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
- 60 s between trials
Conditions
- Ambient
- Not recorded
- Thermal at start
- 33.0 °C, nominal
- Thermal at end
- 37.8 °C, nominal
- Thermal gate
- Android thermal status none (0)
- Throttling detected
- No
Provenance
- Measured by
- Glenn Sonna, Xybrid
- Operator affiliation
- Xybrid
- Run at
- 12 Nov 2026, 12:23 UTC
- Harness
- xybrid-bench 0.4.0 (eb5b63fe15e0e21cb0784eb67836afaa762ee477)
- Methodology
- v0.6
- Publication
- published
- Published
- 16 Nov 2026, 10:23 UTC
- Reviewed by
- Glenn Sonna
Output check
- Result
- Failed
- Method
- sanity-v1-bf16-reference
- Reference
- checks/sanity-v1.json
- Reference SHA-256
- a702dee079525ae0ec90465b1c9e26052c99cb3e1c53e83dea05be4c194fc7ca
- Required passing checks
- 7
- Check arithmetic
- Failed · sha256 4a44dc15364204a80fe80e9039455cc1608281820fe2b24f1e5233ade6af1dd5
- Check capital
- Passed · sha256 5dd272b4f316b776a7b8e3d0894b37e1e42be3d5d3b204b8a5836cc50597a6b1
- Check opposite
- Passed · sha256 f151cd9e12c8994edfa316e4624f956bc68460454869177f287146f8cc956a12
- Check weekday
- Failed · sha256 6a00dfc1dc867e8454c2c8856e1512d9bf02a76710e3411c0972aec886c76c61
- Check color
- Passed · sha256 d486dfbd5fb578340bccbdd0a662527eab38550648d5f44517e5ac71b8824703
- Check sort
- Failed · sha256 0ca7eabde70df147ffea0ec2627afd6c362a0bfc80b48981b8880485cca2360e
- Check animal
- Failed · sha256 d4735e3a265e16eee03f59718b9b5d03019c07d8b6c51f90da3a666eec13ab35
- Check word
- Failed · sha256 f61e68691bdd0f5ebd3252de0396569d624bd2fba240d66fae90edb456a9f48f
- Check copy
- Passed · sha256 d61e6890d0e3571b0c7693612d41f8d059a1fcdef4ba959c0af4144541d41907
- Check needle-start
- Passed · sha256 8b9af2e5b8b4cfddb87d171997b95c471f8d68a68e30fb5177abeb3332372edd
- Check needle-end
- Passed · sha256 e6589c736af32f81bf239c5eda6de41959122ccc1d05463b75e0eb0c70fc831c
- Check continuation-128
- Failed · sha256 7847505d933db2323908d71428d5f0a65999bbad0a763b4574e8cf5a8a469b7a
- Detail
- 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.
Reproduce it
pnpm bench run --plan plans/s4-pixel8-lfm2.5-350m-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 fe35edef675d…