# OpenBMB's MiniCPM5-1B targets on-device AI

> The compact 1B-parameter model brings long-context handling and tool-calling to phones and laptops.

Published by The Open Weights on May 21, 2026. Canonical: https://theopenweights.com/news/minicpm5-1b-qv4h

## Key facts

- Company: OpenBMB
- Model: MiniCPM5-2B
- Version: 5
- Category: Text / LLM
- Modalities: Text / LLM
- License: Other (Open weights, commercial use allowed)
- Parameters: 1B
- Context window: 131K tokens
- Architecture: LlamaForCausalLM
- Precision: BF16
- Weights size: 2.2 GB
- Significance: notable
- Published: 2026-05-21
- Last verified: 2026-07-20
- Hugging Face: https://huggingface.co/openbmb/MiniCPM5-1B
- Canonical URL: https://theopenweights.com/news/minicpm5-1b-qv4h

OpenBMB has released [MiniCPM5-1B](https://huggingface.co/openbmb/MiniCPM5-1B), the latest entry in its MiniCPM line of small language models built to run efficiently on consumer hardware rather than in the data center. At roughly one billion parameters, the model is squarely aimed at on-device deployment, where memory and power budgets are tight.

According to the model's Hugging Face listing, MiniCPM5-1B pairs long-context handling with tool-calling support—two capabilities that push a small model beyond simple text generation and toward practical agentic tasks like retrieval, function invocation, and structured workflows.

## Why it matters

The frontier of open models has largely been a race for scale, but a parallel effort is underway to make capable models small enough to run locally. That approach offers real advantages:

- Lower latency and no per-query cloud cost
- Data that never leaves the device, a plus for privacy
- Offline operation on phones, laptops, and edge hardware

OpenBMB has been among the more consistent contributors to this space, and a dense 1B model with tool-calling reflects a bet that on-device assistants can do useful work without depending on a network connection.

The model is distributed under a custom license, so teams evaluating it for commercial use should review the terms on the [repository](https://huggingface.co/openbmb/MiniCPM5-1B) before shipping. Full specifications, including the exact context window, are available there.

## Get the model

- [Hugging Face](https://huggingface.co/openbmb/MiniCPM5-1B)

## Sources

- [openbmb/MiniCPM5-1B](https://huggingface.co/openbmb/MiniCPM5-1B) — Hugging Face, May 21, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "OpenBMB's MiniCPM5-1B targets on-device AI", The Open Weights, May 21, 2026, https://theopenweights.com/news/minicpm5-1b-qv4h