Apple Releases Open Source Multimodal Machine Learning Model 'Ferret'


On Oct 9, Cloudflare released Clef-omni, an open-weight decision model. It adds audio and full-video support to text and images, handling multimodal content in one API call. Unlike Clef, it natively supports WAV/MP3 and reflects multimodal and architecture upgrades.....
WeChat AI open-sourced WeMM-Embedding, a multimodal embedding model in 2B, 4B, and 9B versions. It supports text, images, video, visual documents, and arbitrary mixed multimodal inputs. Its multimodal isomorphic core enables deep understanding of text and visual information, improving cross-modal retrieval.....
Google launched Gemini Omni1.1 Flash, an AI video model with up to 4K output, upgrading scene expansion, camera control, and efficiency for developers and creators. It supports direct scene extension, pushing multimodal video generation toward higher resolution and longer duration.....
Alibaba Qianwen officially releases the weights of the Max flagship model Qwen3.8-2.4T-A95B for the first time. The model has a total of 2.4 trillion parameters, with 95 billion activated tokens per Token, supporting 260,000 tokens of context natively, extendable to 1,010,000; its architecture continues the advantages of Qwen3.5, excelling in tasks such as programming, office work, research, and long-term intelligence.
NVIDIA has released its first open-source end-to-end full-duplex speech dialogue model, NemotronLabs VoiceChat 11B, which uses a single network to directly perform streaming speech understanding and generation, replacing the traditional ASR, LLM, TTS step-by-step architecture, achieving breakthroughs in real-time speech interaction.