AIDC-AI/Ovis2-34B

Image-Text-to-Text • 35B • Updated Aug 15, 2025 • 4.32k • 151

Note MLLM - Apache 2.0 - 1/2/4/6/18/34B https://huggingface.co/collections/AIDC-AI/ovis2-67ab36c7e497429034874464

FunAudioLLM/InspireMusic-Base

0.5B • Updated Apr 15, 2025 • 22 • 16

Note Audio/Music/Song - Apache2.0 - 0.5/1.5B https://huggingface.co/FunAudioLLM

stepfun-ai/Step-Audio-Chat

Audio-Text-to-Text • 132B • Updated Feb 17, 2025 • 135 • 458

Note MLLM - Apache 2.0 https://huggingface.co/collections/stepfun-ai/step-audio-67b33accf45735bb21131b0b

stepfun-ai/Step-Audio-TTS-3B

Text-to-Speech • 4B • Updated Feb 17, 2025 • 197 • 193

Note TTS - Apache2.0 By StepFun

stepfun-ai/stepvideo-t2v

Text-to-Video • Updated Feb 19, 2025 • 48 • 471

Note Video - MIT - 30B

Note Powerful Chinese open-source distillation dataset (R1) with 110K samples

moonshotai/Moonlight-16B-A3B-Instruct

Text Generation • 16B • Updated Mar 3, 2025 • 29.4k • 186

Note First open MoE by Moonshot AI

Wan-AI/Wan2.1-T2V-14B

Text-to-Video • Updated Mar 12, 2025 • 33.5k • • 1.45k

Note Video - Apache2.0 - 1.3/14B https://huggingface.co/Wan-AI

qihoo360/TinyR1-32B-Preview

Text Generation • 33B • Updated Sep 24, 2025 • 68 • • 329

Note Reasoning model - Apache2.0 - Outperforms the 70B model Deepseek-R1-Distill-Llama-70B and nearly matches the full R1 model in math.

GSAI-ML/LLaDA-8B-Instruct

Text Generation • 8B • Updated Oct 21, 2025 • 220k • 338

Note Fully trained from scratch, LLaDA delivers performance on par with LLaMA3 8B