OpenBMB
3 models across 1 family · Latest: MiniCPM-V 4.6 (2026-05)
Efficient open-source language models for edge devices.
OpenBMB's portfolio covers 3 active models across 1 current family, spanning rag, agents, and long context. Open a model detail page to compare provider routes and sourced benchmarks.
Covers 5 workload areas across 3 active tracked models; last verified 2026-07-07.
Use it for
- Teams evaluating rag, agents, and long context across this lab's releases
- Comparing model families before committing to a flagship
- Migration and pricing follow-ups across 3 tracked models
Do not use it for
- Choosing a hosting provider without opening a model page for price ladders
Active models
3
Current models from this lab, excluding deprecated ones
Active families
1
Current model families from this lab
Open catalog
3 open
3 open source / 0 open weights
Lowest output price
Not tracked
No provider output pricing linked yet
Latest dated release
2026-05-11
MiniCPM-V 4.6
Freshness
2026-07-07
Researched 20d ago
Information
Release cadence
Showing 3 recent dated releases (full timeline below). Latest: MiniCPM-V 4.6 (2026-05-11).
Where this lab wins
- RAG: 1 tracked model with ruler / needle retrieval benchmarks.
- Agentic: 1 tracked model with BFCL, tau-bench, and SWE-bench tool-use coverage.
- Long-context: 1 tracked model with context-token or InfiniteBench-class signal.
- Vision: 1 tracked model with multimodal benchmark coverage.
Flagship quality / price signal
Flagship: MiniCPM-4 8B (best sourced coding quality-per-dollar in this portfolio).
Quality-per-dollar unavailable for this flagship — benchmark coverage or output token pricing is still missing.
OpenBMB is a Chinese AI research organization founded in 2022. Efficient open-source language models for edge devices. OpenBMB ships 1 model family totaling 3 models, with the most recent release MiniCPM-V 4.6 in 2026-05. Notable families include MiniCPM. Use it as a stable reference for lab background, release coverage, and follow-up model pages as they are added. Researchers and evaluators can scan. View official API endpoints, benchmark performance, and coding/agent fit for every OpenBMB model.
About
OpenBMB (Open Lab for Big Model Base), co-founded by Tsinghua NLP researchers and ModelBest Inc. (面壁智能), develops the MiniCPM series of ultra-efficient language models optimized for on-device deployment. MiniCPM4 achieves 3x generation speedup on reasoning tasks, while the multimodal MiniCPM-o 4.5 approaches Gemini 2.5 Flash performance with only 9B parameters. The models are freely available for academic research and commercial use.
Featured models
| Model | Released | Context | Input price ($/1M) | Output price ($/1M) | License | Openness |
|---|---|---|---|---|---|---|
| MiniCPM-V 4.6 | 2026-05-11 | 262k | - | - | Apache 2.0 | Open source |
| MiniCPM-4 8B | 2025-05-01 | 32k | - | - | Apache 2.0 | Open source |
| MiniCPM 2B | 2024-02-01 | 4k | - | - | Apache 2.0 | Open source |
Model families
Recent releases
- MiniCPM-V 4.6- 2026-05-11
- MiniCPM-4 8B- 2025-05-01
- MiniCPM 2B- 2024-02-01
Top comparisons
FAQ
Who founded OpenBMB and when?
OpenBMB was founded in 2022 and is associated with Beijing, China.
What models has OpenBMB released?
OpenBMB ships 3 models across 1 family: MiniCPM.
Is OpenBMB's technology open source?
All tracked models are released under Apache 2.0.
Where is OpenBMB headquartered?
OpenBMB is headquartered in Beijing, China.
What is OpenBMB known for?
Efficient open-source language models for edge devices. Its most prominent tracked family is MiniCPM.
How can I access OpenBMB's models?
OpenBMB's provider availability is tracked on model pages as API and hosting data is verified.
Explore related pages
Last reviewed: 2026-07-07. Data sourced from public lab announcements and provider documentation.