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).
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.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 30.1 | 32.7 | 36.3 |
| Agent Workflow | 44.7 | 49.1 | 55.6 |
| Code Generation | 0.0 | 2.2 | 5.9 |
| Role Play & Narrative | 54.2 | 54.5 | 54.7 |
| Research & Analysis | 21.5 | 25.0 | 29.3 |
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.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Intel(R) Arc(TM) 140V GPU (16GB) | llama.cpp | — | 79.9 tok/s | 78.9 tok/s | 600 MB | 8.192 tokens | 32.7 | 3 |
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.