Boards / Raspberry Pi 5
Raspberry Pi 5 (CPU): LLM benchmarks
The default maker board. LLMs run on the four Cortex-A76 CPU cores with llama.cpp or Ollama.
At a glance
- Measured runs34
- Models tested17
- Fastest run11.73 tok/s
- Qwen2.5 1.5B Instruct, Ollama default, Ollama · source ↗
- Sources8
Where to buy
- Raspberry Pi 5 8GB: $175 official list price after Apr 2026 increase, source ↗ (2026-04-01) Check price on Amazon
- Raspberry Pi 5 16GB: $305 official list price after Apr 2026 increase, source ↗ (2026-04-01) Check price on Amazon
Every run on this board
| Model | Quant | Runtime | Gen tok/s | Prompt tok/s | TTFT ms | Source |
|---|---|---|---|---|---|---|
| Qwen2.5 1.5B Instruct | Ollama default | Ollama | 11.73 | – | – | source ↗ 2026-01-20 |
| Gemma 3 1B | Q4_K_M | Ollama | 11.53 | – | – | source ↗ accessed 2026-10-17 |
| Qwen2 1.5B | Ollama default | Ollama | 11.29 | – | – | source ↗ 2026-01-20 |
| Llama 3.2 1B Instruct | Q4_K_M | Ollama | 11.22 | – | – | source ↗ accessed 2026-10-17 |
| Qwen2.5-Coder 1.5B | Ollama default | Ollama | 10.26 | – | – | source ↗ 2026-01-20 |
| Qwen2.5 1.5B | Q4_K_M | Ollama | 9.97 | – | – | source ↗ accessed 2026-10-17 |
| DeepSeek-R1-Distill-Qwen 1.5B | Ollama default | Ollama | 9.04 | – | – | source ↗ 2026-01-20 |
| Llama 3.2 1B | not reported | CPU backend (per paper) | 8.4 | – | 985 | source ↗ published 2026 |
| SmolLM2 1.7B Instruct | Q4_K_M | Ollama | 8.23 | – | – | source ↗ accessed 2026-10-17 |
| BitNet b1.58 2B-4T | Q1.5 | Ollama | 8.13 | – | – | source ↗ accessed 2026-10-17 |
| Qwen3 1.7B | not reported | CPU backend (per paper) | 7.8 | – | 1653 | source ↗ published 2026 |
| Llama 3.2 3B Instruct | Q40 | distributed-llama v0.12.3 | 5.95 | 14.3 | – | source ↗ 2025-02-15 |
| Granite 3.1 Dense 2B | Q4_K_M | Ollama | 5.81 | – | – | source ↗ accessed 2026-10-17 |
| Qwen2.5 3B | Q4_K_M | Ollama | 5.2 | – | – | source ↗ accessed 2026-10-17 |
| Llama 3.2 3B | Ollama default | Ollama | 4.78 | – | – | source ↗ 2026-01-20 |
| Llama 3.2 3B | Q4_K_M | Ollama | 4.69 | – | – | source ↗ accessed 2026-10-17 |
| Llama 3.2 3B | Ollama default | Ollama | 4.61 | – | – | source ↗ 2025-01-09 |
| Llama 3.2 3B | Ollama default | Ollama | 4.51 | – | – | source ↗ 2025-01-09 |
| Vinci Piccolo 1.0 (Qwen3.5-4B based) | Q4_K_M | llama.cpp | 3.21 | – | – | source ↗ 2026-10-05 |
| Vinci Piccolo 1.0 (Qwen3.5-4B based) | Q4_K_M | llama.cpp | 3.19 | – | – | source ↗ 2026-10-05 |
| Vinci Piccolo 1.0 (Qwen3.5-4B based) | Q4_K_M | llama.cpp | 3 | – | – | source ↗ 2026-10-05 |
| Vinci Piccolo 1.0 (Qwen3.5-4B based) | Q4_K_M | llama.cpp | 2.95 | – | – | source ↗ 2026-10-05 |
| Llama 3.1 8B | Ollama default | Ollama | 2.05 | – | – | source ↗ 2025-01-09 |
| Llama 3.1 8B | Ollama default | Ollama | 1.99 | – | – | source ↗ 2025-01-09 |
| Qwen2.5 0.5B Instruct | not reported | Ollama | 1.54 | – | 8210 | source ↗ accessed 2026-10-17 |
| Llama 2 13B | Ollama default | Ollama | 1.35 | – | – | source ↗ 2025-01-09 |
| Qwen2.5 0.5B | not reported | Ollama | 0.86 | – | – | source ↗ 2026-04-24 |
| Qwen3-VL 2B Instruct | not reported | Ollama | 0.42 | – | 24909 | source ↗ accessed 2026-10-17 |
| Llama 3.2 1B | not reported | Ollama | 0.39 | – | – | source ↗ 2026-04-24 |
| DeepSeek-R1-Distill-Qwen 1.5B | not reported | Ollama | 0.32 | – | – | source ↗ 2026-04-24 |
| Qwen2.5 1.5B Instruct | not reported | Ollama | 0.3 | – | – | source ↗ 2026-04-24 |
| Qwen2.5-Coder 1.5B | not reported | Ollama | 0.3 | – | – | source ↗ 2026-04-24 |
| Qwen2.5 1.5B | not reported | Ollama | 0.27 | – | – | source ↗ 2026-04-24 |
| Llama 3.2 3B | not reported | Ollama | 0.14 | – | – | source ↗ 2026-04-24 |
Can the Raspberry Pi 5 run…
Can it run BitNet b1.58 2B-4T?Can it run DeepSeek-R1-Distill-Qwen 1.5B?Can it run Gemma 3 1B?Can it run Granite 3.1 Dense 2B?Can it run Llama 2 13B?Can it run Llama 3.1 8B?Can it run Llama 3.2 1B?Can it run Llama 3.2 3B?Can it run Qwen2 1.5B?Can it run Qwen2.5 0.5B?Can it run Qwen2.5 1.5B?Can it run Qwen2.5 3B?Can it run Qwen2.5-Coder 1.5B?Can it run Qwen3 1.7B?Can it run Qwen3-VL 2B?Can it run SmolLM2 1.7B?Can it run Vinci Piccolo 1.0?
Have this board? Submit your own run →