LFM3-1.7B-GGUF
Recommended sampling: temperature 0.1, topk 50, repeat penalty 1.1. For math and science: temperature 1.0, topp 0.95, topk 20.
Params
1.7 B
Context
—
Downloads 30d
283 K
Likes
268
Download history
daily snapshots · 9 days283 K150 K
Nov 2Nov 5Nov 8Nov 10
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| LFM3-1.7B-QAD-Q4_0.gguf | Q4_0 | 1.0 GB | 1.9 GB | ✅ Runs comfortably |
| LFM3-1.7B-Q4_K_M.gguf | Q4_K_M | 1.1 GB | 2.0 GB | ✅ Runs comfortably |
| LFM3-1.7B-Q5_K_M.gguf | Q5_K_M | 1.3 GB | 2.1 GB | ✅ Runs comfortably |
| LFM3-1.7B-Q6_K.gguf | Q6_K | 1.4 GB | 2.3 GB | ✅ Runs comfortably |
| LFM3-1.7B-Q8_0.gguf | Q8_0 | 1.8 GB | 2.8 GB | ✅ Runs comfortably |
| LFM3-1.7B-BF16.gguf | GGUF | 3.5 GB | 4.6 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ ollama run lfm3-1-7b-gguf # pin the quantization explicitly $ ollama run lfm3-1-7b-gguf-q4_k_m
$ huggingface-cli download LiquidAI/LFM3-1.7B-GGUF-GGUF \
LFM3-1.7B-Q4_K_M.gguf --local-dir .
$ llama-cli -m LFM3-1.7B-Q4_K_M.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/lfm3-1-7b-gguf
{
"hf_id": "LiquidAI/LFM3-1.7B-GGUF",
"params_b": 1.70,
"context_length": null,
"license": { "id": "other", "commercial": "unknown" },
"downloads_30d": 283113,
"vram_estimates": [
{ "quant": "GGUF", "gb": 4.6 },
{ "quant": "Q4_K_M", "gb": 2.0 }
],
"updated_at": "2026-11-04T01:00:12Z"
}
Specifications
- Parameters
- 1.7 B
- Licence
- other
- First seen on the Hub
- 2026-10-24
- Training datasets
- undisclosed
- Added to our catalog
- 2026-11-02
Compare with any text-generation model
Popular comparisons
Appears in