AI News

A vulnerability costs more to fix the later it is discovered. As coding agents generate code faster than developers can review it, software needs a first line of defense screening each new function in real time. LLM agents have recently become capable of finding real-world vulnerabilities by working as a security expert would: they run analysis tools, explore the surrounding code, test hypotheses and weigh the evidence before forming their own judgement. This reasoning, however, spans many model calls, so that each verdict takes minutes of compute, far too long for real-time screening. Faster detectors, which fine-tune deep learning models or prompt an LLM for a direct verdict, are less accurate: they judge a function without the context an agent gathers, and an LLM's generated verdict is unreliable even when given some of it.
We present REVA (REal-time Vulnerability Analysis), an AI vulnerability detector that brings the security context and code understanding components behind agentic detectors into a single forward pass of an LLM. For each new function, REVA gives the LLM a pre-computed threat model of its module, lightweight static analysis findings and call-graph information. REVA leverages a trained probe to read the LLM's internal representations of the new function in this context. On PrimeVul's held-out test set, REVA on GLM-5.3-Flash reaches 60.0% accuracy in about one second per function. This matches VulTriage (59.2%), the state-of-the-art agentic detector, run on the same LLM, while being over 250 times faster and less than a tenth as costly. On the same prompt, the probe is 14% to 17% more accurate than the LLM's own direct verdict. Across three open LLMs, REVA is at least as accurate as VulTriage, and about 10% more accurate on two of them. An LLM's internal representations thus encode much of what agentic detectors obtain through reasoning.

The financing gives SpaceX another way to fund its rapidly expanding AI ambitions without relying solely on traditional debt markets.

GPT-6.1 Terra: delivering fast and capable computer use at a lower price.

GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.

Shipping fast with AI? Specsight turns code changes into a living product map, feature knowledge, and clear release updates. Keep Product, Sales, and Support in sync with what your software actually does and what changed.

Moli by Lexmount is our browser engine built from scratch for AI agents. It starts faster and uses less memory than Chrome. Use Moli through Lexmount Browser, our cloud browser platform. Run multiple isolated browser sessions with proxy support while we handle the infrastructure. Let your agents browse, fill out forms, and extract data. Try it in our Playground or connect through our SDK.

Humans reason about the physical world by flexibly combining different forms of inference as a problem unfolds. Achieving this flexibility is a major open challenge in AI: capabilities often remain separated across specialized models, limiting how they can be composed and jointly adapted to a task. We introduce $Ψ$-0.6, a unified vision-language world model that adaptively composes perception, language, geometry, and simulation within a shared reasoning process we call streams of thought. These streams interleave images, text, and explicit representations of geometry and motion, all of which the model can generate and interpret through a shared token interface. $Ψ$-0.6 surpasses leading open-weight models zero-shot on multi-view spatial reasoning, with further gains from reasoning through geometric intermediates. It also supports counterfactual simulation: generating and interpreting imagined scenes to determine, for example, whether moving an object would obstruct a pathway. Reinforcement learning from task feedback jointly optimizes the selection, generation, and use of intermediate representations, improving visual question-answering accuracy with gains that transfer to unseen benchmarks. $Ψ$-0.6 offers a path toward adaptive physical reasoning -- not only predicting the physical world but also choosing the representations through which to reason.

Long horizon agentic tasks are now pushing LLM agents past their context window. To continue the task, the agent scaffold has to free up the context window, a process known as context compaction. Compaction is generally lossy by design: agents can silently lose task constraints, instructions, or their reasoning behind earlier decisions. Although compaction is a standard component of coding agent scaffolds, it has not been studied systematically with frontier models on realistic, long horizon agentic tasks. Previous work evaluate smaller open weight models on toy or synthetic trajectories. We fix the agent scaffold (OpenCode), implement five compaction strategies inside it, and evaluate three frontier models on three long horizon benchmarks (PaperBench, ProgramBench, PostTrainBench). For the two 1M context models, the compaction threshold does not measurably affect the task performance, whereas GPT-5.3-Codex, the model with the smallest context window, performs worse at lower thresholds on PaperBench and, less clearly, on PostTrainBench. No compaction method outperforms default self-summarization, and methods that discard more content cost some models up to 20 points while leaving other models the same. Compaction can, however, silently remove task constraints: on ProgramBench, the first compaction summary already drops about one third of the task's anticheating rules on average, and this rises monotonically with the number of compactions. On real user-agent sessions from SWE-chat, instructions about how the agent should work and what it should not do are dropped around 6 times more than specifications of what to build. We argue that current benchmarks are compaction-friendly: i.e., the objective is fixed and the agent's progress stays on disk, so most of what compaction discards can be recovered, except constraints that don't exist anywhere else.

This report introduces Kolibri, an English-German mixture-of-experts transformer with 78.1B total parameters and 3.46B active parameters per token, released as open weights under the Apache 2.0 licence. We built Kolibri for sovereign and specialised deployment, with a particular focus on German and on regulated domains such as public administration, industry, and aerospace. To support these settings, we control the full model-development process, including data, architecture, training infrastructure, post-training, and evaluation.
We train Kolibri on 24T tokens across pre-training, mid-training, and long-context extension, with German accounting for more than 20% of the data mix by bytes. To reach this share, we complement German datasets with more than 2T German tokens that we curate from the web with our own pipeline and generate synthetically. Post-training combines supervised fine-tuning on curated German and English data with reinforcement learning on more than 1.2M internally curated tasks, which cover reasoning, tool use, instruction following, code, and retrieval. The model activates 4.4% of its parameters per token, and among the open models evaluated in this report, it lies on the Pareto frontier of quality and serving cost in English and German.
Kolibri builds on Kolibri Origin, an earlier iteration that validated our full pipeline at scale. Both models come from our Model Factory, which connects data, experimentation, training, evaluation, and deployment through version-controlled workflows.

Nous Research (@NousResearch) on X

It sits in your MacBook's notch. Every Claude Code session gets a card, every sub-agent gets a row, and each one has a little pixel programmer. You see who's working, on which model, and who's waiting on you. Free and open source.

zCAD. AI is an AI-native platform for generating parametric 3D models, built for designers, makers and engineers. Describe a part or assembly in plain language, and zCAD. AI generates an editable, dimension-driven 3D model built on real CAD geometry, not a static mesh. • Complete assemblies from a single prompt, with mates between parts • Refine in plain language, or edit dimensions in a browser-based CAD editor • Export to STEP for use in any CAD tool

Vision-language models (VLMs) can perform a broad range of tasks across diverse settings, yet their performance in agentic contexts remains poorly understood. Although online evaluations in simulators and real-world environments are essential for measuring interactive agent performance, they are costly and challenging to scale. Complementary offline VQA benchmarks can enable faster and more reliable comparisons across agentic domains, yet existing benchmarks remain limited and fragmented. To address this gap, we introduce AgentVQA, an offline benchmark for evaluating VLMs on realistic agentic scenarios across five agent domains: Web Agents, Robotics, Egocentric Videos, Games, and Spatial Planning. We build AgentVQA by manually selecting 1000 challenging, real-world agent examples from 23 source datasets and adapting them for reliable offline evaluation. The resulting benchmark preserves realistic scenarios from interactive environments across diverse domains, and its model rankings track those from corresponding online evaluations. Across the 22 evaluated models, the best achieves only 58.5% accuracy. Our cross-domain analysis shows that errors are driven primarily by failures to ground and maintain accurate representations of the current and evolving visual state, including localizing relevant UI elements, interpreting motion, tracking state changes over time, and inferring 3D spatial relationships, rather than by task misinterpretation or downstream logical reasoning.

Cinematic video generation requires both temporal inference from limited inputs (i.e., sparse rather than frame-by-frame guidance) and factor-aware control of character, scene, action, and camera. Multimodal large language models (MLLMs) offer useful priors for both capabilities, but their representations are not automatically actionable for video generation: a Video Diffusion Transformer (DiT) may bypass them through native conditioning pathways, and using them does not ensure either factor-aware control or generation-compatible temporal prediction. We propose Factor-Aware Predictive Learning (FAPL), which establishes effective MLLM latent use as a shared prerequisite and then targets the two capabilities with complementary learning signals. First, latent-only generator grounding trains the DiT with a frozen MLLM and connector, strengthening the DiT's use of MLLM latents for generation. Second, for factor-aware control, paired videos specify a target-factor change and non-target preservation, and factor-span replacement trains the DiT to realize this relationship in its output. Third, for temporal inference, text-only and text-plus-complete-video observations share a video target, and the trained DiT is then frozen while the MLLM and connector are optimized through it. This final stage trains limited-input representations for generation while retaining complete-observation reconstruction as a complementary training mode. Extensive experiments demonstrate our consistently better performance on three cinematic video generation tasks (i.e., sparse-to-video, factor-aware editing, and multi-shot generation) over baselines and validate FAPL's contributions to MLLM feature use, temporal control, and factor coordination.

Meta filed for $35 billion of new stock offerings and plans a $60 billion at-the-market program, with Nvidia, AMD and Broadcom signaling interest in up to $5 billion of shares.

Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO.

Read more: https://t.co/ofhv0DK3Zp

Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, no…

Llama Windows brings llama.cpp to the Windows 11 system tray. Download open models from Hugging Face in one click, chat via Alt+Space, and let coding agents, editors and scripts share a local OpenAI-compatible API and model cache. Free and open source.

Every classic WordArt style, shape and 3-D effect, in your browser.

Built for Grok Bot, Muse AI, Instinct, Hermes, Claude, GPT, OpenClaw, Cursor, Codex, all agents are welcome. Work on the same live artifacts. Docs, reports, slides, apps, and designs, with shared context, controlled access, and full version history.

Highly recommended read on looped and depth-grown models.

There is a lot of interest in looped and depth-grown language models for reasoning.

The question here is what reusing layers does to in-context learning.

At 3B parameters, both grown and looped models gain more from later context than a d…

On-policy self-distillation enables post-training foundation models without relying on stronger external teachers. However, existing approaches rely on ground-truth question-answer pairs, human-curated problem sets, or other supervised signals. In this work, we reformulate on-policy self-distillation as a self-training framework, driven by self-generated question-answer pairs and privileged reasoning demonstrations. We push supervision to the limit: guided solely by coarse, high-level topic descriptions, the model synthesizes its own curriculum of queries and privileged reasoning traces without access to any human-curated prompts or answers. We find that self-distillation on fully synthetic privileged information enables a model to improve across multiple domains while retaining its capabilities on domains it is not self-trained on. We demonstrate the generality of this framework across text and multimodal domains, achieving consistent self-improvement on standard language benchmarks and Visual Question Answering. Our findings demonstrate that foundation models possess rich latent capabilities that can be amplified without external task supervision or curated problem distributions.

NVIDIA's GeForce RTX 30-series and RTX 40-series graphics cards are based on the same dies as their professional-grade GPUs. This means the compute potential of consumer-oriented RTX GPUs is restricted by NVIDIA's firmware stack, but modders have managed to unlock the instruction issuin

sideOS lives in your Mac’s notch. Hold Control + Option, say what you want, let go! It reads your screen, opens apps, clicks, types and tells you when it’s done. Run it fully offline local models or use your own Claude, OpenAI, Gemini or Grok key.

MyPip turns your game ideas into playable games you can publish and monetize. Describe what you want to build, and AI brings it to life directly in your browser, no coding required. Create multiplayer experiences, add in-app purchases and ads, and take your games beyond the browser to reach more players. From your first idea to publishing and monetization, MyPip gives anyone the tools to become a game creator.

As LLM agents operate over corpora of more than ten thousand tools, selecting the right tool for a complex multi-step user instruction becomes a central bottleneck. Tool descriptions cannot all fit in context, and relevance-based retrievers only rank tools without knowing which are actually needed. Recent work instead leverages proprietary commercial LLMs like Claude or ChatGPT to drive a retriever, decomposing the instruction and selecting over retrieved candidates. We show that a small, open model can be trained via reinforcement learning to perform the same functionality. On ToolBench, a Qwen3-8B model post-trained with our recipe outperforms Llama-4-Maverick (402B). In addition, the trained model transfers directly to different retrievers and generalizes to new tool corpora in different domains.

Generating 360-degree panoramas requires both local perspective fidelity and global spherical continuity. Existing representations face a trade-off: equirectangular projection (ERP) enables seamless panorama modeling but suffers from severe distortion and a domain gap to pre-trained image generators, whereas cubemaps preserve perspective geometry but often produce inconsistent faces and visible seams. We present Squaring-the-Sphere, a panoramic image generation framework built on a novel representation CubeAtlas, which unfolds all cubemap faces into a single compact canvas. By arranging orientation-aligned, rotated polar regions alongside the equatorial faces, this representation preserves the perspective geometry of cubemaps while organizing the cross-face relationships required for seamless panorama assembly. It further enables a pretrained diffusion transformer to model the complete panorama jointly in a single pass, without altering the backbone architecture or positional encoding. While CubeAtlas has a token count comparable to ERP, its structured layout makes many dense attention interactions unnecessary. We therefore introduce a layout-aware sparse attention mask that preserves the interactions required for panorama continuity while removing geometrically irrelevant token pairs. As a result, CubeAtlas achieves better quality than ERP modeling with lower attention computation. Furthermore, we introduce a training-free super-resolution technique and a fusion scheme that enhance fine details in distortion-free perspective space before achieving the final ERP panorama. Experiments on various benchmarks demonstrate that our method improves geometric fidelity, cross-view consistency, and high-resolution visual quality over previous methods.