Model benchmarks

unsloth/LFM2-1.2B-GGUF local LLM performance

As of September 2026, unsloth/LFM2-1.2B-GGUF runs at up to 79.9 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).

llama.cpp
ShareRedditX

Model size

1.2B

Peak speed

79.9 tok/s

Average speed

78.9 tok/s

Avg PP

800.4 tok/s

Min memory

600 MB

Max context

8.192 tokens

Avg output / run

6.812 tokens

Avg runtime / run

1m 28s

Avg quality

32.7

Benchmark runs

3

GPUs tested

1

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 3 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

TaskP5 (low)AvgP95 (high)
Overall30.132.736.3
Agent Workflow44.749.155.6
Code Generation0.02.25.9
Role Play & Narrative54.254.554.7
Research & Analysis21.525.029.3

Performance by hardware and tool

Every hardware/tool/quantization combination unsloth/LFM2-1.2B-GGUF has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Intel(R) Arc(TM) 140V GPU (16GB)llama.cpp79.9 tok/s78.9 tok/s600 MB8.192 tokens32.73

Benchmark runs

All 3 unsloth/LFM2-1.2B-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

Is unsloth/LFM2-1.2B-GGUF good for coding?
In our benchmarks, unsloth/LFM2-1.2B-GGUF scores 2.2/100 for coding. It runs at about 78.9 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
Is unsloth/LFM2-1.2B-GGUF good for agentic (tool-using) tasks?
In our benchmarks, unsloth/LFM2-1.2B-GGUF scores 49.1/100 for agentic workflows. It runs at about 78.9 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
How fast is unsloth/LFM2-1.2B-GGUF for local inference?
Across 3 community benchmark runs, unsloth/LFM2-1.2B-GGUF reaches up to 79.9 tok/s and averages 78.9 tok/s, with the fastest results on Intel(R) Arc(TM) 140V GPU (16GB).
How much memory does unsloth/LFM2-1.2B-GGUF need?
The leanest observed configuration used about 600 MB of memory.
Which tools have been used to run unsloth/LFM2-1.2B-GGUF?
Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.