{"v":1,"id":"model:hf:LiquidAI/LFM2.5-1.2B-Instruct-GGUF","slug":"model-liquidai-lfm2-5-1-2b-instruct-gguf","kind":"model","category":"llm","title":"LFM2.5-1.2B-Instruct-GGUF","summary":"LFM2.5 is a new family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with extended pre-training and reinforcement learning.","source":{"provider":"hf","ref":"LiquidAI/LFM2.5-1.2B-Instruct-GGUF","url":"https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct-GGUF","rev":"8ed288026e23958ad9dfa92d53ed773a8eee7125","fetchedAt":"2026-10-02T21:00:57.580Z","etag":"W/\"1bd3-9D6rq2HoF+8scNz89bEKOfF+Zrc\""},"author":{"name":"LiquidAI","url":"https://huggingface.co/LiquidAI"},"license":{"spdx":null,"raw":"other","open":null,"note":"custom license: read it at the source before installing"},"metrics":{"downloads":350989,"downloadsWeek":327838,"likes":224,"takenAt":"2026-10-02T21:00:57.580Z"},"tags":["safetensors","gguf","liquid","lfm2.5","edge","llama.cpp","text-generation","en","ar","zh","fr","de","ja","ko","es","endpoints_compatible","conversational"],"pipeline":"text-generation","links":{"github":"ggml-org/llama.cpp","npm":"node-llama-cpp"},"updatedAt":"2026-09-22T20:41:54.000Z","collectedAt":"2026-10-02T21:00:57.580Z","review":{"numbers":["350,989 downloads on Hugging Face","224 likes","license other","0.7 GB for LFM2.5-1.2B-Instruct-Q4_K_M.gguf","327,838 npm downloads a week for node-llama-cpp","latest node-llama-cpp@3.22.1"],"log":null},"trust":"unlabeled","health":{"status":"alive","checkedAt":"2026-10-02T21:00:57.580Z","http":200},"description":"Try LFM • Docs • LEAP • Discord\n\n# LFM2.5-1.2B-Instruct\n\nLFM2.5 is a new family of hybrid models designed for **on-device deployment**. It builds on the LFM2 architecture with extended pre-training and reinforcement learning.\n\nFind more details in the original model card: https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct\n\n## 🏃 How to run LFM2.5\n\nExample usage with [llama.cpp](https://github.com/ggml-org/llama.cpp):\n\n```\nllama-cli -hf LiquidAI/LFM2.5-1.2B-Instruct-GGUF --conversation \\\n    --temp 0.1 --top-k 50 --repeat-penalty 1.05\n```\n\n## QAD Q4_0 GGUF\n\nThe Quantization-Aware Distillation (QAD) checkpoint is available as\n[`LFM2.5-1.2B-Instruct-QAD-Q4_0.gguf`](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct-GGUF/blob/main/LFM2.5-1.2B-Instruct-QAD-Q4_0.gguf).\n\nThis is distinct from the post-training-quantized `LFM2.5-1.2B-Instruct-Q4_0.gguf`;\nboth use the GGUF Q4_0 format.\n\nExample usage with [llama.cpp](https://github.com/ggml-org/llama.cpp):\n\n```csharp\nllama-cli -hf LiquidAI/LFM2.5-1.2B-Instruct-GGUF \\\n  --hf-file LFM2.5-1.2B-Instruct-QAD-Q4_0.gguf \\\n  -p \"What is C. elegans?\"\n```\n\n## QAD source weights (safetensors)\n\nThe original FP32 QAD source checkpoint is available in\n[`qad/`](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct-GGUF/tree/main/qad), with its model config,\ntokenizer, generation defaults, and the same chat template as the released QAD GGUF.\nIt can be loaded in Transformers by passing `subfolder=\"qad\"`:\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\nrepo_id = \"LiquidAI/LFM2.5-1.2B-Instruct-GGUF\"\ntokenizer = AutoTokenizer.from_pretrained(repo_id, subfolder=\"qad\")\nmodel = AutoModelForCausalLM.from_pretrained(\n    repo_id, subfolder=\"qad\", dtype=\"auto\", device_map=\"auto\"\n)\n\ninputs = tokenizer.apply_chat_template(\n    [{\"role\": \"user\", \"content\": \"What is 2 + 2?\"}],\n    tokenize=True, add_generation_prompt=True, return_dict=True, return_tensors=\"pt\"…\n\nSource: https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct-GGUF","install":{"kind":"model","hfId":"LiquidAI/LFM2.5-1.2B-Instruct-GGUF","gated":false,"format":"gguf","files":[{"name":"LFM2.5-1.2B-Instruct-BF16.gguf","size":2343326528,"quant":"BF16","sha256":"3d80914b903cd6f3cc041208cf20ec46a3224f840c732e5fd7698832b4743d1b"},{"name":"LFM2.5-1.2B-Instruct-F16.gguf","size":2343326528,"quant":"F16","sha256":"1e1d8a5ca01c0f1ee51a6fd729c80efd626f54812a1241358bea20824fea790d"},{"name":"LFM2.5-1.2B-Instruct-Q4_0.gguf","size":695751488,"quant":"Q4_0","sha256":"2ea801949d760cdf1a2cc04a54262c22c3c0c54f0769d57760c9adeb0e59233f"},{"name":"LFM2.5-1.2B-Instruct-Q4_K_M.gguf","size":730895168,"quant":"Q4_K_M","sha256":"b1b3de114215d9507409a662a501a631095a479a419584e8a2ded6304b19b4f5"},{"name":"LFM2.5-1.2B-Instruct-Q5_K_M.gguf","size":843354944,"quant":"Q5_K_M","sha256":"fa03f3ac4da941a53a0cd4450aacf6a80804c6a1ff885d2fdcbe9406c03215c4"},{"name":"LFM2.5-1.2B-Instruct-Q6_K.gguf","size":962843456,"quant":"Q6_K","sha256":"c5e895c191a066f6b26a8f09f10e94cdb799e579216f87df61a7e27beacd9a2b"},{"name":"LFM2.5-1.2B-Instruct-Q8_0.gguf","size":1246253888,"quant":"Q8_0","sha256":"f6b981dcb86917fa463f78a362320bd5e2dc45445df147287eedb85e5a30d26a"},{"name":"LFM2.5-1.2B-Instruct-QAD-Q4_0.gguf","size":695755488,"quant":"Q4_0","sha256":"bb741ebb106d543e9de114b843a3d3d73d51c74b5801e69da2abde821a0cb3e1"},{"name":"qad/model.safetensors","size":4681379080,"sha256":"992954801538bcb9462f129df78b5ff961ee598148977bb8fbf401d1e35e13ec"}],"totalBytes":14542886568,"suggestedFile":"LFM2.5-1.2B-Instruct-Q4_K_M.gguf","requirements":{"ramGb":2,"diskBytes":730895168,"note":"estimate: suggested file size × 1.15 + 0.5 GB; a real measurement comes with lsh models"},"runWith":["llama.cpp","ollama"],"command":"lsh models install hf:LiquidAI/LFM2.5-1.2B-Instruct-GGUF"}}