Auto Added by WPeMatico
Zhipu AI has open sourced the GLM-4.6V series as a pair of vision language models that treat images, video and tools as first class inputs for agents, not as afterthoughts bolted on top of text. Model lineup and context length The series has 2 models. GLM-4.6V is a 106B parameter foundation model for cloud and […]
Cisco and Splunk have introduced the Cisco Time Series Model, a univariate zero shot time series foundation model designed for observability and security metrics. It is released as an open weight checkpoint on Hugging Face under an Apache 2.0 license, and it targets forecasting workloads without task specific fine tuning. The model extends TimesFM 2.0
How do you keep RAG systems accurate and efficient when every query tries to stuff thousands of tokens into the context window and the retriever and generator are still optimized as 2 separate, disconnected systems? A team of researchers from Apple and University of Edinburgh released CLaRa, Continuous Latent Reasoning, (CLaRa-7B-Base, CLaRa-7B-Instruct and CLaRa-7B-E2E) a
A new computer modeling tool developed by an MIT Energy Initiative (MITEI) research team will help infrastructure planners working in the electricity and other energy-intensive sectors better predict and prepare for future needs and conditions as they develop plans for power generation capacity, transmission lines, and other necessary infrastructure. The tool could reduce the amount
Helping power-system planners prepare for an unknown future Read More »
DeepSeek har lanserat sin senaste AI-modell DeepSeek V3.2 som introducerar flera betydande förbättringar jämfört med tidigare versioner. Modellen bygger vidare på den experimentella version som släpptes i september och påstås prestera på samma nivå som OpenAI:s GPT-5 i flera resonemangstest. En annan stor nyhet är ”thinking in tool-use”, V3.2 kan nu integrera analytiskt tänkande direkt i verktygsanvändning.
Black Forest Labs har nyligen släppt FLUX.2 en avancerad AI-modell för bildgenerering och redigering som erbjuder högupplösta fotorealistiska. Den största nyheten är multi-reference-funktionen som låter dig mata in upp till 10 referensbilder samtidigt. Upplösningen har också fått en boost upp till 4 megapixels vilket ger dig produktionsklara bilder direkt utan behov av upscaling. Ljussättning, fysik
FLUX.2 AI-bildgenerering med upp till 4MP upplösning Read More »
At SAS, we use and contribute to a wide range of open source projects. This series – SAS Contributes – highlights how our teams give back to the open source community. In this installment, we’re focusing on OpenSearch. If you’ve ever sifted through thousands of system logs or product reviews to […] The post SAS contributes:
Liquid AI och Apollo: Din privata AI i fickan Tänk dig att ha en AI-assistent som fungerar helt utan internet och som aldrig skickar din data någonstans. Det är precis vad Liquid AI har skapat med sina nya Foundation Models (LFMs) och Apollo-appen. Den senaste generationen LFM2 är byggd från grunden för att fungera lokalt
Why do current audio AI models often perform worse when they generate longer reasoning instead of grounding their decisions in the actual sound. StepFun research team releases Step-Audio-R1, a new audio LLM designed for test time compute scaling, address this failure mode by showing that the accuracy drop with chain of thought is not an
How can an AI system learn to pick the right model or tool for each step of a task instead of always relying on one large model for everything? NVIDIA researchers release ToolOrchestra, a novel method for training a small language model to act as the orchestrator- the ‘brain’ of a heterogeneous tool-use agent https://arxiv.org/pdf/2511.21689
