---
id: "2026-10-07-openbmb-minicpm-v-4-7"
url: "https://postcutoff.com/e/2026-10-07-openbmb-minicpm-v-4-7/"
as_of: "2026-10-14T00:22:00+02:00"
date: "2026-10-07"
date_precision: day
category: open-source
importance: 2
confidence: medium
status: [Partly confirmed]
sources: 5
editor: human editor
human_review: null
version: "2026-10-13"
---

As of: 2026-10-14 00:22 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-10-07-openbmb-minicpm-v-4-7/

# OpenBMB releases MiniCPM-V 4.7

Full title: OpenBMB releases MiniCPM-V 4.7: an open (Apache 2.0) 35B mixture-of-experts vision-language model with ~3B active parameters, said to run on an iPhone in 3 GB of memory

Around Oct 6-7, 2026 OpenBMB, the open-source group linked to Tsinghua University and ModelBest, published MiniCPM-V 4.7 on Hugging Face under Apache 2.0: a 35B-parameter mixture-of-experts image and video understanding model with about 3B active parameters (35B-A3B), plus a 1B version. The model card claims 13+ tokens per second in 3 GB of memory on an iPhone 17 Pro Max and a higher overall score in less time than Qwen3.6 35B-A3B. Scores are only in chart images; we found no independent tests.

## Key facts
- Repositories (Hugging Face API creation times): GGUF builds of both sizes on Oct 6, 2026; openbmb/MiniCPM-V-4.7-35B-A3B on Oct 7; FP8 and DSpark draft models on Oct 8; int4, AWQ and GPTQ quantizations on Oct 6-9. No blog post or paper was found; the date here is that of the main repository
- Architecture (model card): 'the first MoE model in the MiniCPM-V series'; 35B total parameters, about 3B active; language backbone credited to Qwen3.5 and Qwen3.6, vision encoder SigLIP2, with 'the intra-ViT early compression technique from LLaVA-UHD v4' and 'mixed 4x/16x visual token compression'
- Inputs: single images, multiple images and video; output: text. License: Apache 2.0
- On-device claim (vendor): 'a decoding speed of 13+ tokens/s with just 3 GB of memory on iPhone 17 Pro Max', using Lookahead Expert Scheduling (loading only the experts about to be needed), layer-wise quantization and DSpark speculative sampling with a small draft model
- Platforms named: iOS, Android, HarmonyOS, and 'day-0 support for chip platforms from Rockchip, Houmo, Intel, AMD, and Arm'
- Benchmark claim (vendor, no numbers in text): in Instruct mode the 35B-A3B model 'achieves a higher overall score in less inference time than Qwen3.6 35B-A3B'
- Runs with transformers (>=5.18), vLLM, SGLang, llama.cpp (GGUF) and Ollama (openbmb/minicpm-v4.7). Not served by any Hugging Face inference provider; no API price
- Uptake by Oct 13 (Hugging Face API): the 35B-A3B GGUF repository about 10,100 downloads, the 1B GGUF about 2,100; the main repository 448 likes

## What happened

OpenBMB's MiniCPM-V line is a family of small open vision-language models built to run on phones. Version 4.7 is its first
mixture-of-experts model: 35B parameters are stored, but only about 3B are active for each token, and a scheduling technique loads
only the experts that are about to be used. That is how OpenBMB explains the claim of running a 35B model in 3 GB of phone memory.
A 1B version and several quantized builds were published alongside.

We read the Hugging Face model card and the Hugging Face API listing. The benchmark comparison with Qwen3.6 35B-A3B is given only
as charts, there is no paper, and we found no independent measurement of the iPhone claim, so confidence is medium.

## Why it matters

If the memory claim holds, it is a notable step for on-device multimodal models: a model of the size class that normally needs a
workstation GPU answering questions about images and video on a phone, under a permissive license.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): 160 days after its cutoff
- Claude Opus 5.5 (training cutoff June 2026): 99 days after its cutoff
- Gemini 4 Argon (training cutoff March 2026): 190 days after its cutoff
- Grok 4.7 (training cutoff May 2026): 129 days after its cutoff

## Sources

1. [Hugging Face: openbmb/MiniCPM-V-4.7-35B-A3B (model card)](https://huggingface.co/openbmb/MiniCPM-V-4.7-35B-A3B) (huggingface.co, official)
2. [Hugging Face: OpenBMB organization (all MiniCPM-V 4.7 variants)](https://huggingface.co/openbmb) (huggingface.co, official)
3. [GitHub: OpenBMB/MiniCPM-V](https://github.com/openbmb/MiniCPM-V) (github.com, code)
4. [GitHub: MiniCPM-V 4.7 documentation](https://github.com/OpenBMB/MiniCPM-o/blob/main/docs/minicpm_v4dot7_en.md) (github.com, docs)
5. [Ollama: openbmb/minicpm-v4.7](https://ollama.com/openbmb/minicpm-v4.7) (ollama.com, code)

## Changes

- 2026-10-13 (filed): Created (primary source is the Hugging Face model card)

## Related

- 2026-08-03: [Alibaba launches Qwen3.8-Max and open-sources the Qwen3.8 family](https://postcutoff.com/e/2026-08-03-alibaba-qwen3-8-max/index.md)
- Models: [MiniCPM-V 4.7 35B-A3B](https://postcutoff.com/m/minicpm-v-4-7-35b-a3b/) (`minicpm-v-4-7-35b-a3b`)