1. Qwen3.8-27B by Qwen
image-text-to-text · transformers · Likes: 11143 · Downloads: 665513 · View on Hugging Face →
Tags: transformers, safetensors, qwen3_5, image-text-to-text, conversational, license:apache-2.0, eval-results
What it is. Qwen3.8-27B is a conversational multimodal model from Qwen that handles image-text-to-text tasks, accepting both images and text inputs to produce text outputs.
Why it's trending. It is trending this week due to very high engagement signals in its metadata, including over 665,000 downloads, more than 11,000 likes, and a trending score of 10,560.
Who should use it. It is well suited for developers and teams building conversational, multimodal applications that need image and text understanding, with Azure deployment and endpoint compatibility available.
2. Qwen3.8-27B-GGUF by unsloth
Model · Likes: 1824 · Downloads: 3561466 · View on Hugging Face →
Tags: gguf, qwen3_5, unsloth, base_model:Qwen/Qwen3.8-27B, base_model:quantized:Qwen/Qwen3.8-27B, license:apache-2.0
What it is. unsloth/Qwen3.8-27B-GGUF is a text-based conversational large language model, released by unsloth as an imatrix GGUF-quantized version of the Qwen/Qwen3.8-27B base model.
Why it's trending. It is notable this week for a high trending score of 1724 alongside roughly 3.56 million downloads and 1,824 likes, reflecting strong adoption of this ready-to-run quantized packaging.
Who should use it. It is aimed at developers and enthusiasts who want to run the Qwen3.8-27B chat model with GGUF-compatible, endpoints-compatible tooling, under the permissive Apache-2.0 license.
3. Qwen3.8-2.4T-A95B by Qwen
text-generation · transformers · Likes: 1066 · Downloads: 11212 · View on Hugging Face →
Tags: transformers, safetensors, qwen3_5_moe_text, text-generation, conversational, license:other, eval-results
What it is. Qwen/Qwen3.8-2.4T-A95B is a large Qwen text-generation model (released in safetensors format with a qwen3_5_moe_text architecture) built for conversational language generation tasks.
Why it's trending. It is trending this week with a high engagement signal in its metadata — 1,066 likes, over 11,000 downloads, and a trending score of 1,019 — alongside eval-results and endpoints_compatible tags suggesting published evaluations and serving readiness.
Who should use it. It is suited to developers and researchers building conversational text-generation applications who want a Qwen-family model with reported evaluation results that is compatible with the transformers library and inference endpoints.
4. MiniMax-Music3 by MiniMaxAI
text-to-audio · diffusers · Likes: 962 · Downloads: 11745 · View on Hugging Face →
Tags: diffusers, safetensors, minimax_music3, music-generation, text-to-music, pytorch, sglang-omni, text-to-audio
What it is. MiniMax-Music3 is a text-to-audio model from MiniMaxAI that generates music from text prompts, built on the Diffusers/PyTorch stack.
Why it's trending. It is trending this week with 962 likes, about 11.7K downloads, and a high trending score of 933, indicating strong recent community interest in its text-to-music capabilities.
Who should use it. Music creators, sound designers, and developers building text-to-music generation applications within the Diffusers and PyTorch ecosystem.
5. LTX-2.5 by Lightricks
image-to-video · diffusion-single-file · Likes: 1224 · Downloads: 503632 · View on Hugging Face →
Tags: diffusion-single-file, image-to-video, text-to-video, video-to-video, image-text-to-video, audio-to-video
What it is. LTX-2.5 is a Lightricks AI model for image-to-video generation that also supports text-to-video, video-to-video, and audio-related generation tasks across nine languages.
Why it's trending. This week it is gaining notable traction, reflected in a trending score of 771, over 503,600 downloads, and 1,224 likes on the Hugging Face Hub.
Who should use it. Video and audio creators, developers, and ComfyUI users looking for a single model covering image-, text-, and audio-driven video and audio synthesis.
6. DeepSeek-V4-Pro-0813 by deepseek-ai
text-generation · transformers · Likes: 602 · Downloads: 30985 · View on Hugging Face →
Tags: transformers, safetensors, deepseek_v4, text-generation, conversational, arxiv:2606.19348, license:mit
What it is. DeepSeek-V4-Pro-0813 is a conversational text-generation model from deepseek-ai built on the transformers library for producing natural language outputs.
Why it's trending. It is drawing attention this week with a trending score of 553, over 30,000 downloads, and more than 600 likes on the Hugging Face Hub.
Who should use it. It is aimed at developers and researchers building conversational assistants, chatbots, and other text-generation applications.
7. Qwen3.8-27B-FP8 by Qwen
image-text-to-text · transformers · Likes: 564 · Downloads: 741011 · View on Hugging Face →
Tags: transformers, safetensors, qwen3_5, image-text-to-text, conversational, base_model:Qwen/Qwen3.8-27B
What it is. Qwen/Qwen3.8-27B-FP8 is a 27-billion-parameter, Apache 2.0-licensed FP8-quantized multimodal base model for image-text-to-text and conversational generation tasks, derived from the Qwen/Qwen3.8-27B base model.
Why it's trending. This model is trending this week due to its strong community traction, evidenced by its 533 trending score, over 740,000 downloads, and 564 likes, as an FP8-quantized variant of the popular Qwen3 27B multimodal base model.
Who should use it. This model is suited for AI developers, researchers, and practitioners looking for an open-licensed, FP8-quantized multimodal model for image-text-to-text and conversational tasks that is compatible with standard deployment endpoints.
8. Muse-Glimmer-30B by meta-models
image-text-to-text · transformers · Likes: 1682 · Downloads: 384097 · View on Hugging Face →
Tags: transformers, safetensors, muse_glimmer, image-text-to-text, conversational, arxiv:2504.13181, arxiv:2602.06036
What it is. Muse-Glimmer-30B is a 30-billion-parameter multimodal language model from the meta-models organization designed for image-text-to-text tasks in a conversational setting.
Why it's trending. It is drawing attention this week with a high trending score of 506, 1,682 likes, and 384,097 downloads, along with published eval-results and links to recent arXiv papers.
Who should use it. It is aimed at developers and researchers building multimodal conversational applications such as visual question answering, image-grounded chat, and vision-language assistants.
9. MiniMax-H3 by MiniMaxAI
image-text-to-video · minimax-h3 · Likes: 4145 · Downloads: 2855539 · View on Hugging Face →
Tags: minimax-h3, diffusers, safetensors, text-to-video, image-to-video, image-text-to-video, video-to-video
What it is. MiniMax-H3 is a multimodal model designed to generate synchronized audio-video content from inputs including text, images, and video, supporting a wide range of cross-modal generation tasks.
Why it's trending. It is trending strongly this week, evidenced by its 4145 likes, more than 2.8 million downloads, and trending score of 487, reflecting high community interest in its multimodal generation capabilities.
Who should use it. It is intended for content creators, AI developers, and researchers building applications that require generating synchronized audio and video from text, image, or video prompts.
10. Qwen3.8-27B-Uncensored-FP8 by orcarouter
image-text-to-text · transformers · Likes: 532 · Downloads: 45465 · View on Hugging Face →
Tags: transformers, safetensors, qwen3_5, image-text-to-text, abliterated, qwen3.8, uncensored, ai-red-team, red-teaming
What it is. Qwen3.8-27B-Uncensored-FP8 is a quantized (FP8) multimodal model derived from Qwen3.8-27B that performs image-text-to-text tasks including conversational reasoning and function calling.
Why it's trending. It has attracted notable engagement this week with a trending score of 469, over 45,000 downloads, and 532 likes, likely driven by interest in its abliterated, red-teaming-oriented uncensored variant.
Who should use it. AI red-teamers, safety researchers, and developers exploring uncensored quantized multimodal capabilities for English and Chinese applications.