OpenBMB logo

MiniCPM

MiniCPMv4 / 4.1Current
byOpenBMBOpenBMB(open source)
Context128K tokens
Training CutoffNot publicly specified in queued source
CategoryLarge Language Model
Max Output32.8K tokens

About MiniCPM

MiniCPM is OpenBMB’s ultra-efficient open language-model family for edge and end-device deployment. The MiniCPM4 and MiniCPM4.1 lines focus on fast local reasoning, while MiniCPM-SALA extends the family toward sparse/linear attention and million-token context research.

Capabilities

textcodereasoninglocal inference

Input Modalities

text

Output Modalities

text

Technical Details

API Identifier
OpenBMB/MiniCPM
Category
Large Language Model
Context Window
128,000 tokens
Max Output Tokens
32,768 tokens

Tags

open-weightedge-ailocal-llmreasoningopenbmb

Benchmarks

Performance scores for MiniCPM across standard benchmarks.

MiniCPM-SALA standard benchmark averageofficial-github-readme · May 2026
76.5
MiniCPM-SALA long-context averageofficial-github-readme · May 2026
39
MiniCPM-SALA 2048K extrapolation scoreofficial-github-readme · May 2026
81.6
MiniCPM4.1 reasoning decoding speedupofficial-github-readme · May 2026
3
MiniCPM4 Jetson AGX Orin decoding speedup vs Qwen3-8Bofficial-github-readme · May 2026
7

Pricing and access

Open-weight GitHub and Hugging Face model family. There is no fixed vendor API price; runtime cost depends on the host, hardware, or inference provider.

OpenBMB logo
OpenBMBopen source

Open-source efficient AI models and agent infrastructure

1 ModelsFounded 2022China
View full profile

AI Models by OpenBMB

AI models from the same organization.

ModelContext WindowPrice (In / Out per M)
VoxCPMCurrent4K-- / --