Can I run it? / Raspberry Pi 5
Best LLMs to run on a Raspberry Pi 5
Updated November 2026
As of November 2026, the best model to run on a Raspberry Pi 5 is LFM3 1.7B by Liquid AI (Small LM Index 61.2), at about 8 tokens/s with 4-bit weights.
The Pi 5 runs models on its CPU. Its memory is slow (about 17 GB/s), so speed, not size, decides what is usable: the 16 GB board fits bigger models but runs them no faster. About 6.5 GB is usable for a model, at 17 GB/s. The figures below assume 4-bit weights and a 4K-token context, and only include models that generate at least 5 tokens per second.
Models that run well on a Raspberry Pi 5
All 20 models that run at 5+ tokens/s, ranked by the Small LM Index.
| # | Model | Index | Memory | Speed |
|---|---|---|---|---|
| 1 | LFM3 1.7B Liquid AI · 4-bit · LFM Open License | 61.2 | 1.6 GB | ≈ 8 tok/s |
| 2 | MiniCPM5 2B OpenBMB (MiniCPM) · 4-bit · Apache-2.0 | 60.9 | 2.3 GB | ≈ 6 tok/s |
| 3 | Ling 3.0 Tiny InclusionAI (Ling) · 4-bit · MIT | 54.9 | 5.6 GB | ≈ 5 tok/s |
| 4 | Spark-X2.6 1.7B SparkLLM (XHToken) · 4-bit · Apache-2.0 | 52.3 | 1.7 GB | ≈ 8 tok/s |
| 5 | Edge0 8B-A1B (preview) Edge0 · 4-bit · Apache-2.0 | 50.1 | 5.6 GB | ≈ 5 tok/s |
| 6 | LFM2.5 2.6B Liquid AI · 4-bit · LFM Open License | 44.7 | 2.3 GB | ≈ 5 tok/s |
| 7 | EXAONE 4.0 1.2B LG AI Research (EXAONE) · 4-bit · EXAONE License (non-commercial) | 42.5 | 1.6 GB | ≈ 11 tok/s |
| 8 | LFM2.5 8B-A1B Liquid AI · 4-bit · LFM Open License | 40.1 | 5.9 GB | ≈ 7 tok/s |
| 9 | Gemma 4 E2B Google · 4-bit · Apache-2.0 | 37.4 | 3.7 GB | ≈ 6 tok/s |
| 10 | Qwen3.5 2B Alibaba (Qwen) · 4-bit · Apache-2.0 | 35.9 | 2.0 GB | ≈ 6 tok/s |
| 11 | LFM2.5 1.2B Thinking Liquid AI · 4-bit · LFM Open License | 34.7 | 1.3 GB | ≈ 12 tok/s |
| 12 | MiniCPM5 1B OpenBMB (MiniCPM) · 4-bit · Apache-2.0 | 29.4 | 1.3 GB | ≈ 13 tok/s |
| 13 | LFM2.5 1.2B Instruct Liquid AI · 4-bit · LFM Open License | 25.4 | 1.3 GB | ≈ 12 tok/s |
| 14 | MiniCPM-V 4.6 OpenBMB (MiniCPM) · 4-bit · Apache-2.0 | 24.7 | 1.4 GB | ≈ 11 tok/s |
| 15 | Gemma 3n E2B Google · 4-bit · Gemma | 20.9 | 4.0 GB | ≈ 7 tok/s |
| 16 | LFM2.5 350M Liquid AI · 4-bit · LFM Open License | 16.6 | 0.8 GB | ≈ 40 tok/s |
| 17 | Qwen3.5 0.8B Alibaba (Qwen) · 4-bit · Apache-2.0 | 14.3 | 1.1 GB | ≈ 16 tok/s |
| 18 | Gemma 3 270M Google · 4-bit · Gemma | 12.8 | 0.7 GB | ≈ 52 tok/s |
| 19 | Gemma 3 1B Google · 4-bit · Gemma | 8.9 | 1.2 GB | ≈ 14 tok/s |
| 20 | Llama 3.2 1B Meta · 4-bit · Llama 3.2 Community | 8.7 | 1.4 GB | ≈ 12 tok/s |
Change the context length or quantization, or see models that only just fit, in Can I run it? for this device.
How to run LFM3 1.7B on a Raspberry Pi 5
- Install Ollama. Use 64-bit Raspberry Pi OS. Ollama installs with one command and runs on the CPU.
curl -fsSL https://ollama.com/install.sh | sh
- Run LFM3 1.7B. Downloads the 4-bit GGUF from Hugging Face and opens a chat.
ollama run hf.co/LiquidAI/LFM3-1.7B-GGUF:Q4_K_M
- Keep it cool. Use the active cooler: under sustained load an uncooled Pi 5 throttles and slows down.
Every model page lists the same commands for its own files. LFM3 1.7B's page also has a memory table for other context lengths.
Questions
What is the best LLM for a Raspberry Pi 5?
As of November 2026, the best model to run on a Raspberry Pi 5 is LFM3 1.7B by Liquid AI (Small LM Index 61.2), at about 8 tokens/s with 4-bit weights.
What is the biggest model that runs on a Raspberry Pi 5?
LFM2.5 8B-A1B (8.5B parameters, 1B active) fits at 4-bit, needing 5.9 GB and generating about 7 tokens/s. Bigger models may fit at lower quality or run more slowly; try them in the Can I run it? tool.
How much memory can a model use on a Raspberry Pi 5?
About 6.5 GB of the 8 GB. The Pi 5 runs models on its CPU. Its memory is slow (about 17 GB/s), so speed, not size, decides what is usable: the 16 GB board fits bigger models but runs them no faster.
How accurate are these speeds?
They are estimates from memory bandwidth (17 GB/s here), which limits how fast a model generates text, with a typical efficiency for this kind of hardware. Expect real speeds within about a third either way. Long prompts take longer to read before the first word appears.