CatalogWantedMirrorsLedgerOperating StatementDocsStudio statusCard system
alt.freeweights.minicpm / MiniCPM-V-4.7-35B-A3B-mxfp8
From: mlx-community <huggingface.co/mlx-community>Newsgroups: alt.freeweights.minicpmSubject:

MiniCPM-V-4.7-35B-A3B-mxfp8

Date: 2026-11-15Commit: d801ba7Signed: 2DC28CC7B26A283B ✓

Image text to text model with 35.2B parameters. Safetensors weights, 38.8 GB in all. Published by mlx-community under Apache-2.0.

35.2B params38.8GBMiniCPM-V-4.7-35B-A3B-mxfp8safetensorspermissive
FactsParams35.2BSize38.8 GBFormatssafetensorsTaskimage text to textLicenseApache-2.0From sourcepermissive
$ fw get mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
# manifest sha256 73ce51ba22e2… · 38.8 GB · 20 files
$ fw verify ./MiniCPM-V-4.7-35B-A3B-mxfp8 mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
From sourceHosted (signed)mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8Last check: 2026-11-15 05:15 UTC
Run at homeGet help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Quantized of MiniCPM-V-4.7-35B-A3B.

Works with

Run23Fine-tune29Convert15Evaluate13Scan13Apps62
OverviewModel cardFilesUse itLicenseProvenanceModel tree

Overview

Size
38.8 GB
Weights
38.8 GB (safetensors)
Parameters
35.2B
Formats
safetensors
Task
image text to text
License
Apache-2.0 (permissive)
Files
20, each with a sha256
Publisher
mlx-community
Added
2026-11-15
Section
alt.freeweights.minicpm
Lineage

Original release from the publisher.

Base model: openbmb/MiniCPM-V-4.7-35B-A3B

Where these bytes come from
Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources
BadgeHostLast checkResult
From sourcehuggingface.co2026-11-15 05:15 UTCpassed

Model card

From the publisher at commit d801ba7. Shown as plain text with links; images open at the publisher.

MiniCPM-V-4.7-35B-A3B-mxfp8#

Converted from openbmb/MiniCPM-V-4.7-35B-A3B at revision 17e651151c161dce119cd0ac57570f60da8a0b40 with MLX-VLM main (e928101a144a3f11a27c7ef8d71b6d889b2d250c, version 0.7.8). Support was added in PR #2499.

MXFP8 weight quantization (mxfp8, group size 32). This is mixed precision: vision modules, visual mergers, embeddings, output heads, and linear attention remain BF16; recurrent A_log tensors retain their source precision. The 35B-A3B expert routers use 8-bit affine quantization.

See the original model card for the architecture, intended uses, and limitations. This conversion retains the source model's apache-2.0 license.

Usage#

Install MLX-VLM from main with MiniCPM-V 4.7 support:

pip install -U git+https://github.com/Blaizzy/mlx-vlm.git@main
mlx_vlm.generate --model mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8 --image path/to/image.jpg --prompt "Describe this image." --max-tokens 200 --temperature 0

The native NumPy/Pillow processor requires neither PyTorch nor remote code. It supports text, images, multiple images, and video. The default visual downsampling is 16x; pass --processor-kwargs '{"downsample_mode": "4x"}' for finer detail.

Reproduce conversion#

mlx_vlm.convert --hf-path openbmb/MiniCPM-V-4.7-35B-A3B --revision 17e651151c161dce119cd0ac57570f60da8a0b40 --mlx-path MiniCPM-V-4.7-35B-A3B-mxfp8 --dtype bfloat16 -q --q-bits 8 --q-group-size 32 --q-mode mxfp8

For the exact converter version, install MLX-VLM at commit e928101a144a3f11a27c7ef8d71b6d889b2d250c. Conversion options, per-module quantization overrides, and weight SHA-256 hashes are recorded in conversion.json.

Validation#

The saved checkpoint passed strict reloading, quantization configuration checks, greedy text generation, and cat-image generation in both 16x and 4x downsampling modes on JarvisLabs RTX PRO 6000 Blackwell (cuda) with MLX 0.32.3. See validation.json for the outputs and environment. These are smoke checks, not a quality benchmark. Floating-point 4-bit formats can differ in output quality from affine quantization.

Files

20 files · 38.8 GBWeights 38.8 GB (safetensors). Sizes and sha256 from the signed manifest.
Hide all 20 files with their sha256
FileSizesha256
.gitattributes1.57 KB34448b82c17d60fec9b65b1f093c115ddbaadc04beb1b0140b6bfed2e012a930
README.md2.62 KB3da3761a8b54e325ab613e0f28a788f591b45cc6499987d1e9e8e09153eefd74
chat_template.jinja7.25 KBc4531cacbf2a0ab65170433d7a4f371d11b7599cf96c04b192e6372432636c70
config.json23.9 KB19e464f7837440dc714c6894859f712105ece1b79cecc4f4df3b81e50c5fc53e
conversion.json10.0 KB0f249051b7bec9cd7f3ee24f84973f15501032f7bfc4dbdc55e426696ed1e661
generation_config.json186 B7fa47217744f0e0918f6cf846e7541e4f322e6bc66917044cf2ad580699216c8
model-00001-of-00008.safetensors5.14 GBb38ec4cb658a1904c11fb925f77cb02a406b51e65de855af276e7314d8349aa9
model-00002-of-00008.safetensors5.36 GBa3ef5ffa3775aeb287feff5cce61cf22fe36b9ff98af838fe6e4d70d997c4581
model-00003-of-00008.safetensors5.34 GB7b87fad9f8330493a581c775ef0c843cc9bd78999a59c31d6e09637d5d4b87e1
model-00004-of-00008.safetensors5.36 GB4de71df0fc154c22fbaa3a252dddd95be48995e4b046676735ebed5ea53ec7aa
model-00005-of-00008.safetensors5.34 GB76b8ec835c1733ae9fb02d9897fe58398fefe260399ac976cea63b12128fffb4
model-00006-of-00008.safetensors5.36 GB494d4f4e6add4567d8d6fd09a7a428ae62e670fde868b8eb5a75158e1428d127
model-00007-of-00008.safetensors5.34 GBc773b1199c0c4eb7f224857b6f740c4b4b37eefb3da8f3370b71bf625d55e15a
model-00008-of-00008.safetensors1.57 GB5a70f6137c25c83f969a4821aa2bc079d8e8d8d35dcc785c5906e3fa48104682
model.safetensors.index.json168 KB2283a54b1f9be6e6932369ff7be5fdbc97bff270c1c08f66f08f39d506505dab
preprocessor_config.json148 Bc64289d04f36e48d9d059e3a0069f54062b2ef6044a807adb448b98449cb54de
processor_config.json764 B8540e97ca17871b46aca8639b05f26db4eb4ac39d271b705c9cdbc3e32abbe07
tokenizer.json20.0 MBe4608c5d7b3f4f30ba83be16a82cdac856b2e7191d40f0a08cbacc76c2a926ee
tokenizer_config.json1.16 KBbe74f538ac586f2daac1ede56d421e915d196fba24171bc8fcd05256e42108bc
validation.json1.60 KB793fbccab6c61df9af961eb2beeb55c4a949eb5421b0a4c777df9c84e00c1b61
VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.
Upstream commitDateContents
d801ba7d14 pinned now2026-11-15 05:15 UTC20 files · 38.8 GB

Use it

$ fw get mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
# manifest sha256 73ce51ba22e2… · 38.8 GB · 20 files
$ fw verify ./MiniCPM-V-4.7-35B-A3B-mxfp8 mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
fw.pyChecks the signature and every sha256, tries each source in order, resumes.
$ python3 fw.py get https://freeweights.allternit.com/manifests/mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8.json --dir ./MiniCPM-V-4.7-35B-A3B-mxfp8 $ python3 fw.py verify ./MiniCPM-V-4.7-35B-A3B-mxfp8 ./mlx-community__MiniCPM-V-4.7-35B-A3B-mxfp8.json
hfThe Hugging Face CLI with one setting. Same repo id.
$ HF_ENDPOINT=https://freeweights.allternit.com hf download mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8 --revision d801ba7d14d99cb6a3d9a5f9395ec0ccea578e88 --local-dir ./MiniCPM-V-4.7-35B-A3B-mxfp8
transformersLoad the weights in Python.
import os os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com" from transformers import AutoModelForImageTextToText, AutoProcessor processor = AutoProcessor.from_pretrained("mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8", revision="d801ba7d14d99cb6a3d9a5f9395ec0ccea578e88") model = AutoModelForImageTextToText.from_pretrained("mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8", revision="d801ba7d14d99cb6a3d9a5f9395ec0ccea578e88", device_map="auto")
vLLMServe an OpenAI-compatible API.
$ HF_ENDPOINT=https://freeweights.allternit.com vllm serve mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8 --revision d801ba7d14d99cb6a3d9a5f9395ec0ccea578e88
Run at home

License

Name
Apache License 2.0
Id
Apache-2.0
Class
Permissive
Redistribution
yes
Listing
Hosted (signed)
Decided
License as declared by the publisher. No person read it. 2026-11-15
Text in
README.md
Text sha256
3da3761a8b54e325ab613e0f28a788f591b45cc6499987d1e9e8e09153eefd74
Terms gate
No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: mlx-community <huggingface.co/mlx-community>Newsgroups: alt.freeweights.minicpmSubject: MiniCPM-V-4.7-35B-A3B-mxfp8Date: 2026-11-15Commit: d801ba7Signed: 2DC28CC7B26A283B ✓
Upstream
mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
Pinned commit
d801ba7d14d99cb6a3d9a5f9395ec0ccea578e88
Manifest
MiniCPM-V-4.7-35B-A3B-mxfp8.json · .minisig
Manifest sha256
73ce51ba22e22e506bda4e7c076b5f9e8c355a6cf2afaefd3a90b9d72353e817
Signature
minisign, Ed25519, key 2DC28CC7B26A283B
Created
2026-11-15 05:15 UTC
Generator
freeweights-manifest/0.1.0
Takedowns
None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8.json $ curl -O https://freeweights.allternit.com/manifests/mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8.json.minisig $ minisign -Vm MiniCPM-V-4.7-35B-A3B-mxfp8.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq $ shasum -a 256 <file> # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

Parent

openbmb/MiniCPM-V-4.7-35B-A3B · this model is a quantization of it

No model in the catalog names this one as its parent.

Get it
Download
$ HF_ENDPOINT=https://freeweights.allternit.com hf download mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8
Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest
minisign, key 2DC28CC7B26A283B. MiniCPM-V-4.7-35B-A3B-mxfp8.json. Provenance. Takedown history: none.
Source
huggingface.co/mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp8

Related models

All of alt.freeweights.minicpm →
From: openbmb <huggingface.co/openbmb>Subject: AgentCPM-Explore-GGUF
From: openbmb <huggingface.co/openbmb>Subject: AgentCPM-Report
From: openbmb <huggingface.co/openbmb>Subject: AgentCPM-Report-GGUF
From: openbmb <huggingface.co/openbmb>Subject: BitCPM-CANN-0.5B-gguf
From: openbmb <huggingface.co/openbmb>Subject: BitCPM-CANN-1B-gguf
From: openbmb <huggingface.co/openbmb>Subject: BitCPM-CANN-3B-gguf