Tag

Qwen2.5-Coder is the latest open-source model in Alibaba's Qwen series, focused on tasks such as code generation, inference, and repair.

CogSound is a sound effects generation model developed by Zhipu Technology, designed to create sound effects that match the visual content of AI-generated videos. It integrates closely with the latest video generation model, CogVideoX v1.5, which has achieved significant improvements in video generation capabilities.

CogVideoX v1.5 is the latest open-source video generation model developed by the Zhipu AI team, designed to enhance image-to-video (I2V) generation quality and capabilities.

Hunyuan-Large is Tencent’s recently open-sourced, large-scale Mixture of Experts (MoE) model, featuring 3.89 trillion total parameters and 52 billion active parameters.

Hunyuan3D-1.0 is a recently open-sourced, high-efficiency 3D generation model from Tencent, supporting both Text-to-3D and Image-to-3D generation.

SmolLM2 is a series of compact language models recently released by Hugging Face, designed specifically for on-device applications.

MobileLLM is a highly efficient language model launched by Meta, specifically designed for mobile devices and resource-constrained environments.

Sonic is a low-latency voice generation model developed by Cartesia AI, designed to provide real-time conversational AI solutions.

Stable Diffusion 3.5 marks a significant advancement in the series, showcasing enhanced capabilities for professional-level image generation.

GLM-4-Voice is an advanced end-to-end speech model developed by Zhipu AI, designed to facilitate real-time speech interaction in both Chinese and English. This model features multiple advanced capabilities, including the ability to understand and generate speech, while adjusting emotional tone, pitch, speed, and accent based on user instructions.

Aya Expanse, developed by Cohere For AI, is an advanced multilingual large language model designed to bridge the gap between artificial intelligence and language. The model supports 23 languages and is available in two versions: 8B (8 billion parameters) and 32B (32 billion parameters).

Mochi 1, the latest open-source video generation model released by Genmo, marks a significant advancement in video generation technology.