LFM2.5-24B-A2B-GGUF
LFM2 is a new generation of hybrid models developed by Liquid AI, specifically designed for edge AI and on-device deployment. It sets a new standard in terms of quality, speed, and memory efficiency.
Params
24.0 B
Context
131,072
Downloads 30d
177 K
Likes
239
Download history
tracking started — chart appears after 7 days of snapshots (4 recorded)177 K downloads in the last 30 days
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 |
|---|---|---|---|---|
| LFM2.5-24B-A2B-QAD-Q4_0.gguf | Q4_0 | 13.6 GB | 19.0 GB | ✅ Runs comfortably |
| LFM2.5-24B-A2B-Q4_K_M.gguf | Q4_K_M | 14.5 GB | 20.1 GB | ⚠️ Tight — reduce context |
| LFM2.5-24B-A2B-Q5_K_M.gguf | Q5_K_M | 17.0 GB | 22.8 GB | ⚠️ Tight — reduce context |
| LFM2.5-24B-A2B-Q6_K.gguf | Q6_K | 19.7 GB | 25.8 GB | ❌ Won’t fit |
| LFM2.5-24B-A2B-Q8_0.gguf | Q8_0 | 25.5 GB | 32.1 GB | ❌ Won’t fit |
| LFM2.5-24B-A2B-BF16.gguf | GGUF | 48.0 GB | 56.9 GB | ❌ Won’t fit |
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 lfm2-5-24b-a2b-gguf # pin the quantization explicitly $ ollama run lfm2-5-24b-a2b-gguf-q4_k_m
$ huggingface-cli download LiquidAI/LFM2.5-24B-A2B-GGUF-GGUF \
LFM2.5-24B-A2B-Q4_K_M.gguf --local-dir .
$ llama-cli -m LFM2.5-24B-A2B-Q4_K_M.gguf \
-c 8192 -ngl 99 -t 8 --color
$ curl -s https://aimodelscomparison.com/api/v1/models/lfm2-5-24b-a2b-gguf
{
"hf_id": "LiquidAI/LFM2.5-24B-A2B-GGUF",
"params_b": 24.00,
"context_length": 131072,
"license": { "id": "other", "commercial": "unknown" },
"downloads_30d": 176551,
"vram_estimates": [
{ "quant": "GGUF", "gb": 56.9 },
{ "quant": "Q4_K_M", "gb": 20.1 }
],
"updated_at": "2026-10-22T01:00:26Z"
}
Specifications
- Parameters
- 24.0 B
- Context length
- 131,072
- Licence
- other
- First seen on the Hub
- 2026-10-09
- Base model
- LFM2.5-24B-A2B
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-22
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model