AI Briefing#multimodal-model
The latest AI and developer news about #multimodal-model, with the original source and a short summary.
The latest AI and developer news about #multimodal-model, with the original source and a short summary.
The model supports text and image inputs for coding, visual understanding, and professional knowledge work.
The open-weight multimodal model features a 1M-token context window and improved performance in coding and agentic tasks.
NASA and IBM have released an open-source multimodal AI model and benchmark for integrating lunar surface data.
The 7B-scale open model SenseNova-Vision has been released, capable of performing various vision tasks without separate heads.
Meta has released Muse Glimmer, a 30B multimodal model for local agents, under the Apache 2.0 license.
Microsoft has unveiled Mage, a family of efficient multimodal models built at a 4B parameter scale.
FLUX 3, a multimodal model that jointly trains on images, video, and audio to understand the physical laws of the world, has been unveiled.
Inkling, a 975B-scale open-weight multimodal model designed for fine-tuning, has been released.
LG AI Research has unveiled the architecture of the EXAONE multimodal model, which understands text and images bidirectionally.
Tencent has unveiled the RxBrain model, which combines language reasoning and visual imagination to plan robot actions.
Meta's multimodal reasoning model Muse Spark 1.1 has launched on AI Gateway, offering features optimized for agentic tasks.
The combination of cheap sensors, robotics, and multimodal models is creating new AI business opportunities that leverage data from the physical world.
Keye-VL-2.0, a 30B-scale multimodal model specialized for long-video understanding and agentic capabilities, has been released.
SenseNova-U1 model weights and InternEvo-based training code have been open-sourced.
Provides a roadmap and model list for Native Multimodal Modeling (NMM), which goes beyond modular combination to jointly process multiple modalities.
ByteDance has unveiled Lance, a 3B-scale unified multimodal model that supports understanding, generation, and editing of images and videos.
ByteDance has open-sourced a 3B-scale multimodal model with image, video, editing, and reasoning capabilities.
ByteDance has released Lance, a 3B model that unifies image/video generation, understanding, and editing.
ByteDance has released Lance, a 3B open-source model that combines image and video understanding, generation, and editing.
It proposes a method of first creating an underdrawing with SVG/HTML to improve the text and number accuracy of image generation models.
The AI inference market is fragmenting by use case.
The 27B dense multimodal Qwen has unveiled coding performance on par with much larger models.
H Company has unveiled Holotron-12B, a computer use agent model that applies a hybrid SSM and attention architecture to maximize inference efficiency.
This analyzes the causes of bias occurring in text-to-image (TTI) models and emphasizes the need for improved methodologies to evaluate it.