LLM Reference

MiniCPM Models by OpenBMB

OpenBMBApache 2.0Open source
3 models2024–2026Up to 262k ctx

Details

ResearcherOpenBMB
LicenseApache 2.0(OSI)
Commercial useCommercial use allowed
Models3
Released2024–2026
Max context262k

Capabilities

Vision1 of 3 models
Multimodal1 of 3 models

About

MiniCPM is a family of compact yet powerful language models developed by OpenBMB, designed for efficient deployment on edge devices including smartphones, PCs, and automotive systems. Despite their small size (typically 1-8B parameters), MiniCPM models achieve performance competitive with much larger models. All models are released as open-weight on HuggingFace.

Current Variants

Use-when guidance is derived from seed capabilities, context, release, and replacement fields.

3 in view

Use when the workload needs 262k context, 1.3B parameters, and multimodal inputs.

2026-05262k context1.3B parametersmultimodal inputs

Use when the workload needs 32k context and 8B parameters.

2025-0532k context8B parameters
MiniCPM 2BCurrent

Use when the workload needs 4k context and 2.4B parameters.

2024-024k context2.4B parameters

Release Timeline

3 release groups
2026-05
1 current
MiniCPM-V 4.6
262k context1.3B parametersmultimodal inputs
Current
2025-05
1 current
MiniCPM-4 8B
32k context8B parameters
Current
2024-02
1 current
MiniCPM 2B
4k context2.4B parameters
Current

Specifications(3 models)

MiniCPM model specifications comparison
ModelReleasedContextParametersVisionMultimodal
MiniCPM-V 4.62026-05262k1.3BYesYes
MiniCPM-4 8B2025-0532k8BNoNo
MiniCPM 2B2024-024k2.4BNoNo

Frequently Asked Questions

What is MiniCPM used for?
MiniCPM is used for multimodal, vision language, and vision and multimodal work. The family description and listed model capabilities point to those workloads as the best fit.
How does MiniCPM compare to Claude 3?
MiniCPM by OpenBMB is strongest where you need multimodal, while Claude 3 by Anthropic is the closest related family to check for vision and multimodal work. MiniCPM has 3 listed variants and reaches up to 262k context, while Claude 3 reaches up to 200k context, so compare the specs and pricing tables before choosing a production model.
Which MiniCPM model should I use?
If price is the main constraint, use the pricing table first because MiniCPM does not have complete provider pricing in the local data. For the most capable/latest local choice, evaluate MiniCPM-V 4.6 with 262k context and multimodal inputs.