WHAT’S NEW

New AI models.
Keep up without the noise.

The latest source listings from makers we track, covering language, images, video, audio, and 3D. Catalog checked Nov 2, 2026.

Sorted by the date a source repository was created, which may precede public release. This is a feed of new listings, not confirmed release announcements. Downloadable weights do not automatically mean an open-source license. Coverage and date meanings.

Latest listings

3,186 dated records

Tencent · Downloadable

DepthCrafter

VisionDepth estimation

Hugging Face lists its task as depth estimation. See the model card for supported inputs and setup.

Source listed
Sep 14, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
other
Compare this model →

StepFun · Downloadable

GOT-OCR2_0

Vision

Understands text and images and responds in text.

Source listed
Sep 12, 2024
Parameters
716M
4-bit weights only
≈ 0.4 GB
License
apache-2.0
Compare this model →

Mistral AI · Downloadable

Pixtral-12B-2409

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
12B
4-bit weights only
≈ 6 GB
License
apache-2.0
Compare this model →

Kyutai · Downloadable

moshiko-mlx-q4

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

moshiko-mlx-q8

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

moshiko-mlx

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
7.7B
4-bit weights only
≈ 3.8 GB
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

moshiko-candle-q8

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

moshiko-candle

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
7.7B
4-bit weights only
≈ 3.8 GB
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

moshiko-pytorch

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Sep 11, 2024
Parameters
7.7B
4-bit weights only
≈ 3.8 GB
License
cc-by-4.0
Compare this model →

Kyutai · Downloadable

mimi

Embeddings

Turns content into numerical representations for semantic search and retrieval.

Source listed
Sep 10, 2024
Parameters
96.2M
4-bit weights only
≈ 0 GB
License
cc-by-4.0
Compare this model →

Fish Audio · Downloadable

fish-speech-1.4

Audio generation

Works with speech or audio. Check the model card for transcription and generation capabilities.

Source listed
Sep 10, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
cc-by-nc-sa-4.0
Compare this model →

DeepSeek · Downloadable

DeepSeek-V2.5

Text generation

Generates text for conversations and writing.

Source listed
Sep 5, 2024
Parameters
235.7B
4-bit weights only
≈ 117.9 GB
License
other
Compare this model →

OpenBMB · Downloadable

MiniCPM3-RAG-LoRA

Text generation

Generates text, according to its published model architecture (MiniCPM3ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.

Source listed
Sep 4, 2024
Parameters
Not reported
4-bit weights only
Not reported
License
Not disclosed
Compare this model →

OpenBMB · Downloadable

MiniCPM-Reranker

Text classification

Hugging Face lists its task as text classification. See the model card for supported inputs and setup.

Source listed
Sep 4, 2024
Parameters
2.7B
4-bit weights only
≈ 1.4 GB
License
Not disclosed
Compare this model →

OpenBMB · Downloadable

MiniCPM-Embedding

Embeddings

Turns content into numerical representations for semantic search and retrieval.

Source listed
Sep 4, 2024
Parameters
2.7B
4-bit weights only
≈ 1.4 GB
License
Not disclosed
Compare this model →

OpenBMB · Downloadable

MiniCPM3-4B

Text generation

Generates text for conversations and writing.

Source listed
Sep 3, 2024
Parameters
4B
4-bit weights only
≈ 2 GB
License
apache-2.0
Compare this model →

BAAI · Downloadable

CCI3-HQ-Classifier

Capabilities unconfirmed

A specialized research model. See the model card for its intended task and supported setup.

Source listed
Aug 30, 2024
Parameters
567.8M
4-bit weights only
≈ 0.3 GB
License
apache-2.0
Compare this model →

Cohere · Cloud

Command R

Text generationTool use

Cloud API for text generation, chat & writing, tool use. Capabilities are catalog-listed, not independently scored.

Source listed
Aug 30, 2024
License
Provider terms
Compare this model →

Cohere · Cloud

Command R+

Text generationTool use

Cloud API for text generation, chat & writing, tool use. Capabilities are catalog-listed, not independently scored.

Source listed
Aug 30, 2024
License
Provider terms
Compare this model →

What makes this feed useful?

Capability tags distinguish a model that reads images from one that creates them. Each listing links to its source and a permanent profile, so you can inspect its license and compare it with a model you already know.

Does this cover every new release?

No. The current feed tracks public Hugging Face metadata from selected makers. It can miss untracked makers, cloud-only releases, and records beyond a source’s retrieval limit. Announced models without a source listing are not included. Check Sources & methodology for the current coverage.

Listings update with the catalog automatically. A missing or delayed source can leave older information visible; each model profile shows when it was checked.