JanusCoder-14B is an open-source foundation model with 14B parameters built on Qwen3-14B, aiming to establish a unified visual programming interface for code intelligence. This model is trained on the JANUSCODE-800K multimodal code corpus and can handle various visual programming tasks in a unified manner.

Natural Language ProcessingTransformers

JanusCoder-8B is an open-source code intelligence foundation model built on Qwen3-8B, aiming to establish a unified visual programming interface. This model is trained on JANUSCODE-800K (the largest multimodal code corpus to date) and can handle various visual programming tasks, including data visualization, interactive web UI, and code-driven animation.

MultimodalTransformers

GLM-4.5-Air is a foundation model for agents launched by Zhipu AI. It adopts a compact design, unifying reasoning, coding, and agent capabilities to meet the complex requirements of agent applications. The model has a moderate number of parameters and superior efficiency. Multiple versions have been open-sourced and can be used for commercial purposes and secondary development.

Natural Language ProcessingTransformersMultiple Languages

LFM2-Audio-1.5B is the first end-to-end audio foundation model launched by Liquid AI, designed specifically for low-latency and real-time conversations. With only 1.5 billion parameters, this model can achieve seamless conversational interactions, and its capabilities can rival those of models with much larger parameter scales.

MultimodalSafetensorsEnglish

TildeOpen LLM is an open-source foundational language model specifically designed to serve underrepresented Nordic and Eastern European languages. Funded by the European Commission and trained on the LUMI supercomputer, this model has 30 billion parameters and addresses the performance gap faced by users of 19 key languages in existing AI systems.

Natural Language ProcessingTransformersMultiple Languages

GLM-4.5V-AWQ-4bit is a quantized multimodal model built on Zhipu AI's next-generation flagship text foundation model. Optimized by AWQ-4bit quantization technology, it significantly reduces the computational resource requirements while maintaining excellent performance. This model achieves SOTA performance among models of the same scale in 42 public visual language benchmark tests and has powerful visual reasoning capabilities.

MultimodalTransformersMultiple Languages

DINOv3 is a series of general visual foundation models developed by Meta AI. Without fine-tuning, it can outperform specialized state-of-the-art models in a wide range of visual tasks. This model uses self-supervised learning to generate high-quality dense features and performs excellently in various tasks such as image classification, segmentation, and depth estimation.

Computer VisionTransformersEnglish

DINOv3 is a versatile visual foundation model developed by Meta AI. It can outperform specialized models in a wide range of visual tasks without fine-tuning. This model can generate high-quality dense features and performs excellently in various visual tasks, significantly surpassing previous self-supervised and weakly supervised foundation models.

Computer VisionTransformersEnglish

DINOv3 is a series of general visual foundation models developed by Meta AI. It can outperform specialized advanced models in various visual tasks without fine-tuning. The model adopts the Vision Transformer architecture and is pre-trained on 1.689 billion web images. It can generate high-quality dense features and performs excellently in tasks such as image classification, segmentation, and retrieval.

Computer VisionTransformersEnglish

Foundation-Sec-8B-Instruct is an 8-billion-parameter instruction-tuned language model specifically designed for network security applications. Based on the Llama-3.1-8B architecture, it has excellent instruction-following capabilities and professional knowledge in network security. It supports local deployment and can help organizations build AI security tools, reducing reliance on cloud services.

Natural Language ProcessingTransformersEnglish

Cosmos-Predict2.5 is a high-performance pre-trained world foundation model suite developed by NVIDIA specifically for physical AI. Based on diffusion model technology, it can generate high-quality images and videos with physical awareness based on text, image, or video input, providing world simulation capabilities for applications such as autonomous driving and robotics.

MultimodalDiffusers

GLM-4.1V-9B-Base is an open-source vision-language foundation model developed by Zhipu AI. It has 9 billion parameters, focuses on multimodal reasoning capabilities, supports both Chinese and English, and can process images with up to 4K resolution and a context length of 64K.

MultimodalTransformersMultiple Languages

Cosmos-Predict2 is a high-performance pre-trained world foundation model designed for physical AI development, capable of generating physics-aware images, videos, and world states.

MultimodalDiffusers

Cosmos-Predict2 is a series of high-performance pre-trained world foundation models designed for physical AI to generate physics-aware images, videos, and world states.

MultimodalDiffusers

Cosmos-Predict2 is a series of high-performance pre-trained world foundation models designed to generate physics-aware images, videos, and world states, which can be used for the development of physics AI.

MultimodalDiffusers

Magma is a foundational multimodal AI agent model capable of processing image and text inputs to generate text outputs, with complex interaction abilities in both virtual and real-world environments.

MultimodalTransformers

Derm Foundation is a machine learning model designed to accelerate AI development for skin image analysis in dermatology applications.

Computer VisionTensorflowEnglish

The CXR Foundation Model is a machine learning model optimized for chest X-ray image analysis, generating embedding vectors through pre-training to accelerate AI development.

MultimodalTensorflowEnglish

TCRT5 is a sequence - to - sequence (seq2seq) pre - trained model based on the T5 architecture, specifically designed to handle T - cell receptor (TCR) and peptide - MHC (pMHC) sequences. It is pre - trained through masked span reconstruction, aiming to learn the complex relationships between TCR and pMHC sequences and provide a foundation for downstream TCR design, de - orphaning, and related classification/regression tasks.

Natural Language ProcessingTransformers

AI21 Jamba 1.5 is a series of advanced foundation models with powerful long context processing capabilities and efficient inference speed, suitable for various business scenarios.

Natural Language ProcessingSafetensorsMultiple Languages