# LFM3-1.7B

> LFM3-1.7B is a text model for local agents and specialized applications, based on our new tri-hybrid architecture.

- **Organisation:** LiquidAI
- **Category:** Chat & Text Generation
- **Parameters:** 1.7B
- **Context length:** 262,144 tokens
- **Licence:** lfm1.0 (Commercial use with conditions)
- **Released:** 2026-10-23
- **Hugging Face:** 103.7K downloads (30 days), 1.8K likes
- **Ollama:** `ollama run lfm3:1.7b`

## Estimated memory

| Quant | ≈ GB |
|---|---|
| Q2_K | 1.43 |
| Q3_K_M | 1.7 |
| Q4_K_M | 1.9 |
| Q5_K_M | 2.08 |
| Q6_K | 2.27 |
| Q8_0 | 2.67 |
| F16 | 4.26 |

## Runs with

- [llama.cpp](https://selfhost.directory/project/llamacpp)
- [LM Studio](https://lmstudio.ai)
- [MLX](https://github.com/ml-explore/mlx-lm)
- [Ollama](https://selfhost.directory/project/ollama)
- [Transformers](https://github.com/huggingface/transformers)
- [Jan](https://jan.ai) (community)
- [KoboldCpp](https://selfhost.directory/project/koboldcpp) (community)
- [LocalAI](https://selfhost.directory/project/localai) (community)
- [ONNX Runtime](https://onnxruntime.ai) (community)

## About

LFM3-1.7B is a text model for local agents and specialized applications, based on our new tri-hybrid architecture.

• Agentic training: post-trained for agentic harnesses such as Hermes Agent, Pi, and OpenClaw, with a 256K-token context window.
• Designed for the edge: at 4-bit precision, it takes about 1 GB of memory, so it fits on any phone or laptop.
• Made to be specialized: adapt the model to your domains, formats, and tools via quick single-GPU fine-tuning.

Find more information about the LFM3 family in our release blog post.

The LFM3 family includes base, text, vision, audio, omni, and decision models.

LFM3-1.7B is designed for local agents, tool calling, structured outputs, and domain-specific fine-tuning. Use retrieval or search for knowledge-heavy tasks. It is not recommended for agentic coding or complex multi-step reasoning due to its small size.

Choose a checkpoint variant for your inference framework and hardware.

Checkpoint Framework Description BF16 vLLM , SGLang , Transformers BF16 weights for GPU inference. FP8 SGLang FP8 quantization for GPU inference. NVFP4 SGLang 4-bit floating-point quantization for GPU inference. GGUF llama.cpp , Ollama , LM Studio Local inference on CPUs and GPUs. MLX MLX 4-bit quantization for Apple Silicon. ONNX ONNX Runtime Cross-platform inference with ONNX Runtime.

Sampling parameters:

## Alternatives

- [openbmb/MiniCPM5-2B](https://selfhost.directory/models/openbmb/minicpm5-2b) (2.5B, Apache 2.0, ≈2.32 GB at Q4): We are releasing MiniCPM5-2B, the second model in the MiniCPM5 series, following MiniCPM5-1B. It is a dense 2B Transformer that scales up the same training recipe, built for on-device, local deployment, and resource-constrained scenarios, reaching 2B-class open-source SOTA.
- [LiquidAI/LFM3-1.7B-Base](https://selfhost.directory/models/liquidai/lfm3-1.7b-base) (1.7B, lfm1.0, ≈1.9 GB at Q4): LFM3-1.7B-Base is the pre-trained checkpoint behind LFM3-1.7B, a text model for local agents and specialized applications based on our new tri-hybrid architecture.
- [Qwen/Qwen3-1.7B](https://selfhost.directory/models/qwen/qwen3-1.7b) (2B, Apache 2.0, ≈2.58 GB at Q4): Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.
- [Qwen/Qwen3-1.7B-Base](https://selfhost.directory/models/qwen/qwen3-1.7b-base) (1.7B, Apache 2.0, ≈2.4 GB at Q4): Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models.
- [Qwen/Qwen2.5-1.5B-Instruct](https://selfhost.directory/models/qwen/qwen2.5-1.5b-instruct) (1.5B, Apache 2.0, ≈1.63 GB at Q4): Qwen2.5 is the latest series of Qwen large language models. For Qwen2.5, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters. Qwen2.5 brings the following improvements upon Qwen2:
- [XHToken/Spark-X2.6-1.7B](https://selfhost.directory/models/xhtoken/spark-x2.6-1.7b) (1.7B, Apache 2.0, ≈1.95 GB at Q4): We introduce Spark-X2.6-1.7B, the second model in the Spark-X2.6 series, following the release of Spark-X2.6-4B. Building on Spark-X2.5-1.7B, this release further advances our compact, general-purpose language models for on-device deployment.

More: https://selfhost.directory/models/liquidai/lfm3-1.7b/alternatives

## Runs well on

- [GeForce RTX 4060](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/rtx-4060): F16, est. 43 tok/s
- [Apple M4 (16 GB)](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/apple-m4-16gb): F16, est. 19 tok/s
- [Radeon RX 6700 XT](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/rx-6700-xt): F16, est. 61 tok/s
- [GeForce RTX 3060 12GB](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/rtx-3060): F16, est. 57 tok/s
- [Arc B580](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/arc-b580): F16, est. 73 tok/s
- [Tesla T4](https://selfhost.directory/models/liquidai/lfm3-1.7b/on/tesla-t4): F16, est. 51 tok/s

## Similar size

- [meta-llama/Llama-3.2-1B-Instruct](https://selfhost.directory/models/meta-llama/llama-3.2-1b-instruct) (1.2B)
- [Qwen/Qwen2.5-1.5B-Instruct](https://selfhost.directory/models/qwen/qwen2.5-1.5b-instruct) (1.5B)
- [Qwen/Qwen3-1.7B](https://selfhost.directory/models/qwen/qwen3-1.7b) (2B)
- [HuggingFaceTB/SmolLM2-1.7B](https://selfhost.directory/models/huggingfacetb/smollm2-1.7b) (1.7B)
- [tencent/Qwen3.8-27B-DSpine-b16](https://selfhost.directory/models/tencent/qwen3.8-27b-dspine-b16) (2B)
- [meta-llama/Llama-3.2-1B](https://selfhost.directory/models/meta-llama/llama-3.2-1b) (1.2B)

## FAQ

**How much VRAM does LFM3-1.7B need?**
About 1.9 GB at Q4_K_M with an 8K context. Q8_0 needs about 2.7 GB and full precision about 4.3 GB. Longer contexts need more memory for the KV cache.

**Can LFM3-1.7B run on the RTX 4060?**
Yes. It fits at FP16 / BF16 (about 4.3 GB), at an estimated 43 tokens/s. It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.

**Can I use LFM3-1.7B commercially?**
Yes, with conditions. It is released under lfm1.0. Custom licence: read the model card before using it commercially.

**What is the context length of LFM3-1.7B?**
262,144 tokens (256K). The full window needs far more memory than an 8K context.

**How do I run LFM3-1.7B with Ollama or llama.cpp?**
With Ollama: ollama run lfm3:1.7b. With llama.cpp: llama-server -hf LiquidAI/LFM3-1.7B-GGUF:Q4_K_M, which starts an OpenAI-compatible server.

## Links

- huggingface: https://huggingface.co/LiquidAI/LFM3-1.7B
- homepage: https://www.liquid.ai/blog/lfm3
- ollama: https://ollama.com/library/lfm3

---
Source: https://selfhost.directory/models/liquidai/lfm3-1.7b