AI Briefing#multimodal
The latest AI and developer news about #multimodal, with the original source and a short summary.
The latest AI and developer news about #multimodal, with the original source and a short summary.
The new d1-3B model achieves a score of 48.57 on the Decision Index 0.2.1, outperforming larger models like Decider 35B-A3B.
The model scores 44.86 on Decision Index 0.2.1, ranking #1 among 4B models, and processes 1M-token documents in ~3 seconds via API.
The 9B multimodal model claims 8 of 9 top scores in its size band on spatial benchmarks and includes an open-sourced data pipeline.
Embed 5 Pro achieves an average score of 85.8 on ViDoRe V3, outperforming competitors like Voyage 4 Large and Gemini Embedding 2.
The 8B parameter model achieves competitive results with Qwen-Image 20B on specific benchmarks while using a single backbone for both generation and understanding.
The new models accept shared state, named questions, and optional images to return answer distributions in a single forward pass.
The new integration supports a 124B parameter sparse MoE model with 5.5B active parameters and a 256K token context window.
The MiMo-V2.6-Pro released by Xiaomi recorded higher intelligence scores compared to similar-sized open-weight models and features competitive pricing, though its output speed was found to be slower than average.
Xiaomi has launched three MiMo V2.6 models supporting coding and multimodal capabilities on AI Gateway.
OpenRouter has officially released the multimodal model 'Unbiased Pareto' after completing stealth testing.
A single-file C-based inference engine has been released that performs text and image inference for Gemma 1/2/3 without external dependencies.
The VisTW benchmark for evaluating VLM performance in Traditional Chinese environments has been released, with Twinkle Eval providing support.
InternLM has released Intern-S2-397B, a 397B-parameter multimodal model designed for scientific problem-solving and long-horizon agent tasks.
The Agnes-3.0-Flash 33B model, released on Hugging Face, supports a hybrid attention architecture and a 262k context window.
inclusionAI has released Ling-3.0-flash-VL, a multimodal model featuring a 1M token context window and MoE architecture.
A solo developer from India has released Arcle V1, an open-weight 5.8-billion-parameter multimodal model based on a single neural network.
PyTorchKR summarized 10 recent AI/ML papers centered on three major trends: computational efficiency, autonomous agents, and multimodal reasoning.
Ling-3.0-flash-VL has been released with visual understanding and agent capabilities.
World Labs has unveiled Atlas, a next-generation world model that integrates text, image, video, and 3D data.
Researchers at Seoul National University have released Sori-1B, a 1B-parameter language model trained solely on audio data without text pre-training.
LangChain released two design approaches and an evaluation benchmark for multimodal RAG for slide decks.
Alibaba has released the Wan3.0 model, capable of generating 30-second videos from documents.
llama.cpp now supports dots3-note, the first open-weight model in the dots3 family.
DeepSeek has released the v4-flash vision API, which processes text and images together.