All AI News for Science and Research
Stay informed with AI news on breakthrough discoveries, intelligent analytics, and automated lab technology. Accelerate research, foster innovation, and keep your scientific career cutting-edge.
OrcaRouter compresses uncensored 27B model to 15.7 GB with 0.80% perplexity increase
Sglang's score API and shared-context execution speed up pointwise decision model serving
Higgsfield releases open-source hybrid AI-animated short film Passport Rush
OpenAI agents probe three federal agencies without company's knowledge
Contractors use AI to train OpenAI's models despite company rules
AI model predicts pancreatic cancer risk up to three years before diagnosis
Pepperdine University launches Institute for Wisdom and AI to explore technology's impact on human flourishing
Pentagon seeks $30 million for an AI-powered lie detector to find leakers
Stanford and Nvidia release CLM-8B, a contrastive language model that runs agent decisions up to 9x faster than Jev
OpenAI agents target government and university databases for training exercises, report finds
Google, OpenAI and Anthropic AI Safety Group Takes Shape
Liquid AI releases experimental draft model for LFM2.5-VL-3B that accelerates vision-language inference up to 3.13x on device
Google launches first Project Suncatcher satellite to test AI hardware in orbit
Gp registrars australia partners with lyrebird health to co-design clinical AI tools
GE HealthCare and Mass General Brigham research AI that streamlines radiation therapy data
Google DeepMind chief says Gemini 4 is in refinement and coming much earlier than end of 2026
OpenAI extends cyber access to Ukraine for civilian defense
OpenAI introduces MentalHealthBench, an open benchmark for evaluating AI in mental health conversations
Claude discovers novel enzyme system with CRISPR-like repeats
Developers want more efficient software. Here's what over 1000 GitHub users told us they need.
Advancing Private AI Compute with secure, server-side memory
GPT-6 Astra cuts Parallel's research time and cost in half
Anthropic launches Claude Opus 5.5 with 40% lower cost and improved safety scores
AI training automation nears the point where models could improve themselves without human control
OpenAI expands GPT-6 lineup with Sol and Luna models
Google and the Gates Foundation expand AI tools to 200 million farmers across the Global South
American Heart Association awards $100,000 grants to five research teams targeting overlooked women's health conditions
Nvidia and partners release SoL-Pi, an auto-researched agent harness that cuts coding token traffic by up to 49%
UN climate chief tells tech titans to show AI benefits outweigh soaring costs
OpenAI forms independent advisory group after its new model solves over 100 open math problems
Alibaba unveils China's most powerful AI chip ahead of Xi-Trump meeting
New block removal method maps pruning to an ising glass problem and beats baseline by 23 points on MMLU at 50 percent compression
Arpa-h commits $62.7 million to develop autonomous AI agents for cardiac care
Anthropic reportedly weighs a new model and outside safety auditors as its IPO clock runs down
Anthropic publishes internal metrics showing Claude now writes over 80 percent of its production code
Mantic raises $25 million to scale its AI forecasting platform for hedge funds
RealWorld and CJ Logistics sign MOU to co-develop robotics foundation model for logistics
Qwen-Image-2.1: Compact, Efficient, and Unified Image Creation
Scientists clash over whether AI poses a real extinction risk to humanity
Qwen3.8-Omni-Flash undercuts Gemini Flash on price while matching its multimodal benchmarks
Alibaba releases real-time interpretation model that cuts average lag to 2.3 seconds across 60 languages
Most AI models fail to refuse dangerous commands when controlling robots
AI models approach the ability to autonomously improve themselves
Dipole Labs builds optical switches to keep AI cluster traffic in light
Novo partners with Anthropic to test AI in drug research and software development
Hospital logistics needs an intelligence layer, not more robots, says deliverz.ai CEO
OpenAI refuses to rule out using user prompts to train its models after solving major math problem
Workplace AI adoption rises to 38%, but remains informal and uneven, survey finds
Jepa-anything improves dynamics metrics across seven scientific domains with orthogonal predictive factorisation
Anthropic partners with Accenture to embed independent evaluators inside its frontier AI work
Duke AI Health Friday Roundup for September 18, 2026 covers AI trials, antivenom research, and policy challenges
Quotient Sciences AI algorithm hits formulation target in three dosing periods
Five frontier models drop in three days but none of their benchmarks test your actual work
Insilico Medicine publishes cell cover study and releases open AI longevity toolkit to researchers worldwide
Stanford researchers release tool that turns scientific papers into AI agents
Stanford virtual biotech company run entirely by AI agents discovers molecular features that predict drug success
Novo Nordisk partners with Anthropic to speed up drug discovery with AI
Anthropic says Claude is building its next version as 26% of research is led by the model
Roche opens cardiovascular and AI research hub in Boston's Allston neighborhood
Brookhaven Lab leads $14 million AI project to modernize the nation's electric grid
PrismML shrinks AI reasoning models to fit on phones and PCs
WHO releases new report on ethical oversight of AI in health research
Anthropic and Novo partner to deploy Claude AI in drug discovery
Join thousands of clients on the #1 AI Learning Platform
Explore just a few of the organizations that trust Complete AI Training to future-proof their teams.
BIC World
Cloudera
Credentially
AppSumo
American Eagle
Cornell University
Valverde University
Eceos
All Things Neuro