How fast is LFM2.5-8B-A1B-Q4_K_M for local inference?

Across 4 community benchmark runs, LFM2.5-8B-A1B-Q4_K_M reaches up to 285.0 tok/s and averages 281.2 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.

How much memory does LFM2.5-8B-A1B-Q4_K_M need?

The leanest observed configuration used about 3.9 GB of memory (quantizations tested: Q4_K_M).

Which tools have been used to run LFM2.5-8B-A1B-Q4_K_M?

Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.

Model benchmarks

LFM2.5-8B-A1B-Q4_K_M local LLM performance

As of September 2026, LFM2.5-8B-A1B-Q4_K_M runs at up to 285.0 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).

llama.cppQ4_K_M

ShareReddit X

Model size

Peak speed

285.0 tok/s

Average speed

281.2 tok/s

Avg PP

1005.8 tok/s

Min memory

3.9 GB

Max context

32.768 tokens

Avg output / run

19.280 tokens

Avg runtime / run

1m 12s

Avg quality

50.4

Benchmark runs

GPUs tested

Quality by task

Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 4 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

Task	P5 (low)	Avg	P95 (high)
Overall	47.0	50.4	54.4
Agent Workflow	69.4	72.6	76.8
Code Generation	4.6	9.1	14.3
Role Play & Narrative	63.6	70.9	78.9
Research & Analysis	31.7	48.8	61.8

Performance by hardware and tool

Every hardware/tool/quantization combination LFM2.5-8B-A1B-Q4_K_M has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
NVIDIA GeForce RTX 4070	llama.cpp	Q4_K_M	285.0 tok/s	281.2 tok/s	3.9 GB	32.768 tokens	50.4	4

Benchmark runs

All 4 LFM2.5-8B-A1B-Q4_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

How fast is LFM2.5-8B-A1B-Q4_K_M for local inference?: Across 4 community benchmark runs, LFM2.5-8B-A1B-Q4_K_M reaches up to 285.0 tok/s and averages 281.2 tok/s, with the fastest results on NVIDIA GeForce RTX 4070.
How much memory does LFM2.5-8B-A1B-Q4_K_M need?: The leanest observed configuration used about 3.9 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run LFM2.5-8B-A1B-Q4_K_M?: Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.