Research & Benchmarks - Page 5 of 28
Academic AI research, performance benchmarks, scientific breakthroughs, and peer-reviewed studies advancing artificial intelligence frontiers.
547 articles
View complete article list
GLM-5.2 beats GPT-5.5 on SWE-bench Pro (62.1 vs 58.6) for 1/6 cost
For one-sixth the cost, GLM-5.2 just punched above its weight on SWE-bench Pro, scoring 62.1 to GPT-5.5’s 58.6. That single number, though, is only the start.
AMD builds Llama 3.1 8B pretraining benchmark for MLPerf, using random weights
AMD has submitted a training benchmark for a model that doesn't learn. For the latest MLPerf results, the company pretrained Meta's Llama 3.1 8B architecture using entirely random weights. No data, no real gradients.
AMD's MI355X CDNA4 GPU Shows Competitive Training Times in MLPerf v6.0
AMD just matched Nvidia’s top chip. In a head-to-head sprint, the company's new MI355X accelerator fine-tuned a Llama 2 70B model in the same time as a Nvidia B200 GPU, using an eight-accelerator setup.
NVIDIA Blackwell Leads MLPerf Training 6.0 with Full‑Stack Scale
Nvidia won everything. The MLPerf Training 6.0 benchmark results are in, and the company's Blackwell platform took first place in every single test. That clean sweep isn't about a chip.
DR-DCI Enables Agent-Callable Retrieval to Expand Local Workspace Efficiently
Most AI research papers sell you a new way to fail. They’ll claim they’ve cracked some fundamental tension, then quietly fudge the data. This one’s different. DR-DCI fixes a real problem.
Fused kernels boost MoE training, forward and backward passes up to 1.3×
Training a large Mixture-of-Experts model often feels like herding cats on a supercomputer. The GPU is constantly starting and stopping tiny tasks, stuck waiting for messages between experts, and never really working at full capacity.
Hybrid Open-Ended Tri-Evolution Improves Deep Research for AI Agents
Deep research is where AI agents usually fail. They can pull up facts, but they can't learn from them. Their knowledge is frozen. Meanwhile, a separate line of work, called agent evolution, has shown real promise.
Microsoft Research Mirage adds persistent spatial memory to video generation
Video generation has long suffered from a quiet, expensive flaw: every time a model renders a new viewpoint, it must rebuild the world from scratch, pixel by pixel, memory bleeding away with each frame.
Amazon security research prompts White House ban on Anthropic Fable
Amazon's security team found a hole. The White House answered the phone. Now some of the best researchers in the country can't touch their own work. Andy Jassy got a call from Washington after his team flagged something in Anthropic's Fable model.
Study: AI coding agents locate correct file but miss key lines in bugs
AI coding assistants are great at finding the file. They're terrible at reading it.
OpenAI confirms cooperation as state attorneys general launch investigation
OpenAI just beat Elon Musk in court. Now it's facing a different kind of fight. State attorneys general are investigating the company, and OpenAI says it's cooperating. It won't say who or what they're asking for. This isn't happening in a vacuum.
Gemini‑SQL2 leads BIRD benchmark with 80.04% execution accuracy
Google's Gemini-SQL2 just hit 80.04% execution accuracy on the BIRD benchmark. That's a specific, hard number. For context, OpenAI's GPT-5.5-xhigh sits at 72.8%. Claude Opus 4.6 is at 70.9%.
NVIDIA tops AA‑AgentPerf benchmark, credits Vera Rubin platform
Leaderboards are usually marketing noise. This one is different. NVIDIA just topped the first major benchmark for AI agent performance, and the margin isn't close.
Perplexity routes deep‑research subtasks across 20+ models using Gemini agent
Most AI search is still just a fancy text predictor. Perplexity decided to build a factory instead. Its new system breaks a single query into pieces and farms them out to over twenty different AI models, all working at once.
Visual model exploits similarity of 打, 拍, 拉; text model starts from embeddings
Being able to see is not the same as being able to read. A new comparison of training methods for Chinese characters proves it. The visual model gets a head start. It recognizes that the characters 打, 拍, and 拉 all share the same hand-shaped radical.
New arXiv Paper Introduces Strategic Decision Support for AI Agents
We used to think of support as something a computer gave a person. Now the person is often the backup for the computer. This is a real problem. An AI agent booking a flight or managing a supply chain doesn't have a bad day.
LSEG integrates trusted data into ChatGPT workflows, says Max Grigoryev
Everyone selling AI says it's trustworthy. Most are hoping you don't check. The London Stock Exchange Group is taking a different, more literal approach: they're feeding their own vetted financial data directly into ChatGPT's machinery.
Hermes Agent Builder Unites Identity, Model, Skills, Servers in One Dashboard
The dashboard is the command center. Hermes Agent’s Profile Builder collapses fragmentation into one coherent pane of glass: identity, model, skills, and MCP servers all flow together.
SciConBench launches with 9.11K questions to test AI scientific synthesis
Forget whether AI can write your emails. The real question is whether it can do science. A new, brutally difficult benchmark called SciConBench makes it clear that the answer, for now, is a hard no.
Language Agents Self‑Gate Clarification: Mandatory vs Opportunistic Modes
Most AI that talks to you is guessing. It’s making a probabilistic wager on what you meant. A new paper suggests the smartest move a language model can make isn't a better guess, but knowing when to stop guessing altogether.
Featured Resources & Reviews
No Code MBA Course Review
Learn to build AI-powered apps without coding. Our comprehensive review of No Code MBA's course.
AI Tools & Resources
Curated collection of AI tools, courses, and frameworks to accelerate your AI journey.
Weekly AI Digest
Get the week's most important AI news delivered to your inbox every week.