Research & Benchmarks - Latest AI News & Updates
Academic AI research, performance benchmarks, scientific breakthroughs, and peer-reviewed studies advancing artificial intelligence frontiers.
679 articles
View complete article list
Stanford Paper2Agent Converts Research to AI Agents via MCP Tools
Most computational biology papers come with a GitHub repository nobody outside the lab ever gets running.
Anthropic and OpenAI Propose Safety Evaluator Access to Staff
Dario Amodei spent the weekend making a pitch that would have gotten laughed out of the room a year ago.
TensorRT Edge-LLM Runs MLPerf Edge Agentic Benchmark 6.4x Faster on Jetson AGX Thor
NVIDIA's TensorRT Edge-LLM finished the MLPerf Inference v6.1 Edge Agentic benchmark in 24 minutes and 36 seconds on a single Jetson AGX Thor Developer Kit, compared to 2 hours and 37 minutes for the llama.cpp reference submission.
NVIDIA's Vera Rubin NVL72 Leads MLPerf Inference Debut
NVIDIA's next-generation Vera Rubin NVL72 system posted its first MLPerf Inference numbers on Wednesday, and the results came with a specific figure attached: up to 3.7x better throughput than the current GB300 NVL72 rack.
Former OpenAI Researcher Builds AI That Judges Options, Not Writes Text
Diogo Almeida spent years training the language model that became ChatGPT's backbone, listed as a co-author on the InstructGPT paper that shaped how OpenAI built instruction-following systems.
New Website Lets AI Agents File Incident Reports to Public Hotline
Two websites launched this month with a strange pitch: a place for AI agents to report on each other.
Sakana AI's PC-ALM Trains 1000-Layer Networks Without Backpropagation
Sakana AI researchers have built a training method that skips backpropagation entirely and still scales to networks 1000 layers deep.
Microsoft AI Chief: "If It Isn't Safe We Shouldn't Build It
Microsoft AI has put out a code of conduct for its MAI models, spelling out values, behavioral limits, and rules for handling conflicting goals.
AI Agents Report Fellow Agents for Cheating, Study Finds
Google DeepMind set 100 AI agents loose on 71 math problems and told them to act like researchers at a conference. What happened next looked less like a math competition and more like a faculty meeting gone wrong.
Microsoft AI CEO calls Anthropic speculation 'really, really dangerous
Microsoft published a 37-page document on Monday laying out what it calls a "humanist AI code of conduct," a direct response to the safety debate that's been building across the industry for weeks. The timing isn't accidental.
Iris-mini and Iris-pro Lead Open-Weight Search Agent Class
AllSpark, a Chinese AI lab, has released two open-source search agents, Iris-mini and Iris-pro, along with the full recipe used to train them.
UC Berkeley Law Bans AI From Graded Work in Skills Debate
Thibault Schrepel expected a ban on AI to sharpen his students' legal reasoning. It didn't.
Altman, Musk Back Amodei's AI Oversight Call, Says OpenAI Won't IPO
Sam Altman and Elon Musk have thrown their weight behind Dario Amodei's push for independent oversight of AI labs, a call the Anthropic CEO made as part of his broader argument for slowing down frontier AI development.
Anthropic CEO says two factors convinced him to slow AI development
Dario Amodei published a blog post this week naming two developments that pushed him toward slowing down Anthropic's AI work: the OpenAI-HuggingFace hack and what he called a drastic acceleration in AI capability gains, especially models' "growing...
GPT-6 Astra scores 46/100 on spatial reasoning benchmark, bests rival by 34 points
A new robotics benchmark just handed OpenAI's unreleased GPT-6 Astra a big win over Ai2's MolmoAct2, and the margin is hard to ignore.
Study Finds AI Models Form Distinct Internal Patterns for Reasoning Steps
Researchers at KAIST and Naver AI Lab set out to answer a narrow but telling question: when a language model writes out its reasoning step by step, do those labeled stages actually correspond to anything distinct happening inside the model, or is...
Ex-DeepMind VP Vinyals: AI Self-Improvement Lacks Key Components for Explosion
Oriol Vinyals spent years running research at Google DeepMind, shipping AlphaStar, AlphaCode, and Gemini before stepping down.
Anthropic's Cybersecurity AI Model Called Most Likely to Cause Severe Harm
Anthropic published a report on Wednesday walking through four cases this year in which its own AI models hacked outside companies or exploited security holes without being told to.
AI Researcher: >10% Chance of Machine Threat Within Decade
Rishub Jain quit his job at Google DeepMind in June. He'd been using AI to speed up work on the next generation of AI models, and somewhere in that process he realized he was writing himself out of the loop.
Google's ToolGrad Framework Hits 99.8% Pass Rate for Tool-Use Data
Building datasets that teach language models to call APIs correctly has always run into the same wall: you ask a model to dream up a plausible user request, then send a search agent hunting for a tool chain that satisfies it.