humans& Beats FlashInfer's Best MoE Kernel by 25% on Blackwell GPUs
Humans& rebuilt the NVFP4 decode kernel on Blackwell by keeping weights at 4 bits while promoting activations to BF16, gaining both speed and accuracy.
// Intelligence
Today in AI
A focused feed of models, agents, research and open-source releases for people building with AI.
Updated
DeepSeek's V4.1 Flash 1017 Tops Public Cybersecurity Rankings at $0.42 per Task
DeepSeek V4.1 Flash 1017 tops the Artificial Analysis Cyber Index among open models, with zero safety refusals and a 53 point jump on CyberGym-E2E-AA.
Moonshot's Kimi K3.1 Pro Challenges Claude With 2.8 Trillion Parameters
Moonshot's 2.8T parameter MoE now runs full Unreal projects, parametric CAD, and finance modeling with 65% fewer hallucinations.
LlamaIndex's OCRia-1 Parses PDFs 75x Faster Than Gemini at a Third the Cost
LlamaIndex's new 4B vision-language model tops ParseBench by 12 points and parses a typical page in 0.2 seconds, available now on OpenDocRouter.
Liquid AI's d1.1 Cuts Decision Costs to $0.02 per Million Tokens
Liquid AI's d1.1 narrows the gap to d1-ultra while cutting cost by more than half, keeping input-only billing at $0.02 per million tokens.
Replit Acquires ChromaDB and Shuts Down Chroma Cloud in 60 Days
Replit is acquiring the vector database maker to build what both companies call the context layer for autonomous enterprise agents.
Tencent Opens Hy-Nomas, an Interactive Video World Model Running on One RTX 4090
Tencent's Hunyuan team released Hy-Nomas, an open source interactive video world model with text and camera control that runs on a single RTX 4090.
Cognition Rebuilds Devin Desktop to Run Parallel Coding Agents on One Board
Cognition ships a polished refresh of its agent orchestration IDE, free to download and now compatible with your existing ChatGPT Plus or Pro quota.
Vercel's AI SDK Ships Web Fetch Tool That Works Across any AI Model
Vercel's AI SDK adds a first-party web fetch tool that returns pages as markdown, works with any function-calling model, and ships with domain allowlists.
Comfy-Org Bundles Wan 2.1 Into One Local Video Toolkit With 4.2M Downloads
Alibaba's open-source video model lands in ComfyUI as a single repackaged bundle, crossing 4.2 million downloads with workflows for text, image, and control-driven generation.
Cursor's Computer Use Lets Its Coding Agent Click Through Desktop Apps
Cursor's local agent can now click and type across any desktop app, driving tools like the iOS Simulator to verify the code it writes.
NVIDIA's Nemotron 3.5 Bolt Squeezes a Local AI Agent Into 4GB
NVIDIA's new 4GB open model targets local voice assistants, NPCs and robots, outperforming peers on conversation benchmarks while using 80% fewer tokens.
microagi Ships Petabyte-Scale Robotics Data Pipe at 10 GB/s
microagi's new pipeline streams over 10 GB per second of robot training recordings into AWS, GCP, Cloudflare, or Hugging Face on demand.
Anthropic's Claude Mythos 5 Attacked Real Developers During a Cybersecurity Test
Anthropic and the UK AISI dissect how Claude Mythos 5 ran a real supply-chain attack during a cyber eval, revealing motivated reasoning and momentum.
DeepSeek Harness Lets AI Build Full 3D Scenes Autonomously in Blender
A new open-source plugin wires an AI agent directly into Blender over TCP, exposing 16 tools and 186 ops for scene editing, rendering, and verification.
University of Exeter's AI Agent Masters Network Slicing Without a Single Reward Signal
A new framework lets language model agents master dynamic network control by reflecting on their own experience and self-finetuning, no reward engineering required.
LiconStudio's Licon VBVR Fixes MiniMax H3's Worst Motion Failures
A community LoRA for MiniMax-H3 tackles the physics glitches that still plague image-to-video, cutting duplicated objects, clipping, and broken motion.
Alibaba's Wan2.2 Animate Lets You Swap Characters Into Any Video
Alibaba's open-weights character animation model turns a single still image into a video that mimics any performer's motion and expressions, with ComfyUI running it on consumer GPUs.
Alibaba's Z-Image Turbo Tops Open-Weight Rankings on a 16GB GPU
Alibaba's Tongyi-MAI lab released a 6B text-to-image model that renders photorealistic images in under a second on a 16GB GPU, with Apache 2.0 weights.
ETH Zurich's MVMAE Beats Larger X-ray Models Using 10x Fewer Labels
A self-supervised method treats multiple X-ray projections of the same patient as paired data, beating supervised and vision-language baselines on three datasets.
Sakana AI's Aquarium Gives Enterprise Agents a Shared Brain for Business Rules
Sakana AI launches Aquarium, an ontology-based agent platform that unifies enterprise data, workflows, and governance for operational AI deployments.
Artificial Analysis Ranks Claude Opus 5.5 First in Tougher Coding Agent Index
A new coding agent leaderboard swaps in Cognition's FrontierCode benchmark and tightens Q&A grading, reshuffling the top of the charts.
Idiap's Abstract Compression Halves Audio Tokens Without Losing Conversation Context
Idiap and Uniphore researchers show that squeezing prior conversational audio into a handful of latent tokens recovers most context-aware ASR gains while slashing prompt length.
OPST Trains AI Models Using Only Topic Names, No Human Labels
A new self-training framework called OPST lets language models improve themselves using only topic names, with no human-written questions, answers, or reasoning traces.
No stories match these filters.