We ran 900 DeepSWE rollouts on GLM-5.3 and GLM-5.3 Flash. Flash gives up 5.6 points of pass@1 at 17x lower cost, and only 2.6 points at pass@4.
Creative teams produce more assets than ever, but fragmented tools and manual context transfer slow production. This post shows how to build a reusable agent harness with Amazon Quick and fal, connected through the Model Context Protocol (MCP), using two hands-on workflows: an eight-panel storyboard and a music-video concept prototype.
Factual question answering (QA) typically assumes a single canonical answer, obscuring whether large language models (LLMs) retain divergent accounts of long-tail facts. To address this gap, we introduce ElephantBench, a closed-book knowledge probe comprising 1,094 questions generated through an auditable graph-based pipeline. The pipeline retrieves related documents from a low-exposure web corpus
Recent generative approaches to geometry estimation adapt pretrained image diffusion models and treat the task as image-conditioned generation. Leveraging off-the-shelf image diffusion models, they either (i) train task-specific geometry models (for depth and surface normal estimation) independently, losing the opportunity of exploring the intrinsic correlation of these geometric targets, or (ii)
An avatar that holds a conversation should decide what to say and to move while saying it, yet these abilities live in separate model families: spoken dialogue models produce speech without motion, and co-speech motion models produce motion only from audio handed to them. The standard remedy is a cascade that first generates the spoken response and then runs a motion model over the finished audio,
Interactive dialogue games test a capability that static benchmarks largely leave implicit: a model must carry state across turns, interpret feedback, and choose valid actions under changing constraints. We study this setting in the LM Playschool Challenge with a 2B open-weight model, and find that many failures are not only broad knowledge failures but also local decision failures: repeated guess
Due to the nature of quadratic attention, Large Language Models (LLMs) consume a lot of memory and energy. Every new token costs more than the previous one. For each additional token, the keys and values must be stored in memory indefinitely, which is unsustainable. Several alternatives have been proposed to fix the quadratic scaling problem, one of which is retrofitting LLMs to use Linear Attenti
Autoregressive video diffusion enables scalable long-video generation by producing chunks from a bounded recent context. While recency-based caching preserves local continuity, it evicts historical cues needed when subjects, objects, scenes, or attributes reappear. Existing memory mechanisms expose models to nonlocal history, but access alone does not ensure effective use. Our analysis reveals tha
Conversational Recommender Systems (CRS) typically require domain-specific dialogue data, which is costly, scarce, and often unavailable in new domains. We conduct a systematic empirical study of zero-data CRS bootstrapping: generating synthetic conversational supervision from non-conversational signals---item reviews, metadata, and user-item interactions---without any in-domain dialogue corpus. W
Spatio-temporal video grounding (STVG) requires models to identify when a referred event occurs and localize the target entity throughout that interval. Existing multimodal large language models typically serialize dense localization trajectories autoregressively, causing decoding latency to grow with tube length and allowing localization errors to propagate across time. We introduce Parallel Tube
Loop Engineering is emerging as a practice for organizing development work around coding agents. Instead of writing each prompt by hand, practitioners design loops that monitor progress, assign work, run checks, and decide what the agent should do next. Even with a capable coding agent, a loop may trust a stale progress note, skip needed verification, spend its budget in the wrong direction, or st
LLM agents increasingly modify their own prompts, tools, middleware, resources, and execution harnesses at runtime. Such self-evolution can improve capability, but a successful mutation may leave persistent effects that cannot be safely reversed in states different from the one in which it was created. We introduce EvoUndo, a framework for representing, synthesizing, diagnosing, and independently
While Large language models (LLMs) incorporate user personalization signals to improve usability and helpfulness, they increasingly shift from providing balanced, informative responses toward optimizing for user satisfaction when conditioned on personal context such as conversation history, inferred preferences, and user profiles. Specifically, we identify three emerging risks: (1) irrelevant pers
Generative models can turn natural-language prompts into images, text, code, and other content, lowering the cost of producing drafts and components. Their practical impact increasingly depends on whether those pieces can become complete, dependable deliverables. This survey examines agentic artifact creation, which we define as stateful construction in which an AI system materially constructs or
Long-horizon agentic tasks require large language models (LLMs) to iteratively retrieve, integrate, and maintain dispersed information across multi-turn interactions, but preserving all interaction histories leads to a continuously growing working context. Recent proactive context management methods allow models to edit their own working context with specialized tools, yet they still face three ke
Quantization has been widely adopted in LLM training and inference to reduce cost and improve efficiency. However, low-bit quantization of the attention module often introduces large errors at very low bit-widths, causing performance degradation. Existing methods mainly rely on smoothing techniques to handle outliers, while we propose a hybrid quantization design to better balance accuracy and eff
Interactive web application generation requires models to produce usable HTML, CSS, and JavaScript applications from natural language requests. Unlike conventional code generation, application quality depends on multiple user-facing functional requirements, each often tied to localized code regions such as event handlers, state updates, DOM fragments, or CSS selectors. Standard GRPO collapses thes
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
Amazon Bedrock now supports the OpenAI GPT-5.6 models, Terra and Luna, in India with India geographic cross-Region inference. If you have local data processing requirements, you can now use these models at scale while Amazon Bedrock keeps inference requests and data within India.
223 points, 67 comments
218 points, 156 comments
Range-Doppler (rD) maps produced by chirp- sequence (CS) radar systems are fundamentally limited in reso- lution by bandwidth, carrier frequency, and coherent processing interval constraints. Improving resolution through hardware is often impractical due to regulatory, cost, and real-time operation requirements. In this work, we investigate deep learning-based super- resolution of rD maps in both
Self-hosted speech AI carries an observability trade-off: the numbers that drive capacity planning and cost management stay locked inside the vendor container. Deepgram closes that gap on Amazon SageMaker AI with two capabilities that land billing, usage, and per-GPU metrics directly in your own Amazon CloudWatch account.
Serving automatic speech recognition (ASR) models at scale is costly when each request uses only a fraction of a GPU. Learn how NVIDIA CUDA Multi-Process Service (MPS) with NVIDIA Triton Inference Server on Amazon EC2 GPU instances cuts GPU infrastructure by 75% while holding sub-second latency at 92.1 requests per second per GPU.
When Pythian rolled out Google Cloud’s
Ultrashort pulse formation from noise represents a fundamental self-organisation process in nonlinear dissipative systems and remains central to ultrafast photonics. Mamyshev oscillators offer a particularly valuable platform for investigating this phenomenon because they do not rely on conventional saturable absorbers. Instead, pulse formation is governed by self-phase modulation in normal-disper
A randomized study of more than 1,000 students examines ChatGPT, critical thinking, originality, and student performance on a real-world university assignment.
OpenAI is expanding its presence in Brazil, deepening engagement with developers, businesses, and communities to support AI adoption across the country.
339 points, 201 comments
August 27, 2026
We’re opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical devices, to a first group of scientific research labs and advanced manufacturers.
Scaling robot data is crucial for building generalist Vision-Language-Action (VLA) models, yet robot trajectories are harder to scale than web-scale image-text data because embodied collection is costly and sparsely covers the physical world. This makes representation quality a central bottleneck: under a fixed robot-data budget, continued pre-training must turn limited trajectories into transfera
Streaming 3D reconstruction from extremely long videos requires estimating camera motion and scene geometry online under bounded memory and computation. Early streaming models achieve causal, bounded-cost inference using finite context buffers or compact recurrent states, yet their estimates often deteriorate as sequences grow. Recent methods improve long-horizon stability by coupling short-range
Multimodal GUI agents have emerged as a promising paradigm for digital task automation, yet transitioning from benchmark-oriented models to dependable real-world applications remains challenging due to limited environment coverage, brittle task construction, and unreliable reward verification. In this work, we present UI-Venus-2, a general-purpose foundation GUI agent designed to operate across mo
Image retrieval has traditionally been formulated as a point-wise matching problem, where each candidate image is scored in isolation. However, this atomic paradigm fails to capture the complexity of human search intent within personal photo collections, where users often seek compact visual stories bound by structural relations rather than isolated snapshots. To address this limitation, we introd
Recurrent fast-weight memories and selective state-space models compress an expanding context into a fixed-size recurrent state, making the state transition an online learning rule. We study this rule under read-after-write autoregressive semantics. For the prefix-prediction objective considered here, the local fast-memory example revealed at step t is the prefix-aligned pair (x_t,y_t)=(ϕ(k_{t-1})
Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs). However, these methods typically rely on repeated generation or external verification. To address this limitation, we introduce CritICL, a novel inference-time framework that improves reasoning while maintaining high efficiency. Our key insight is that LLM failure modes
Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers such skills from agent experience, which enables agents to progressively adapt through interaction. However, the insights that guide skill development typically remain scattered across optimization histories, limiting their systematic reuse across
Recent prominent post-training methods, such as Reinforcement Learning (RL) and On-Policy Self-Distillation (OPSD), have driven rapid progress in mathematical reasoning for large language models, yet their reliance on ground-truth labels precludes test-time training (TTT). Replacing ground truth with majority-vote pseudo-labels is a natural alternative, yet it is fragile: an incorrect vote corrupt
Multimodal large language models (MLLMs) can interpret a street view, but urban agency depends on whether such local evidence remains useful after the agent starts to move. In this paper, we investigate how far current MLLM agents can turn local urban perception into reliable action in a complicated real-scale city. We propose UrbanGround, the first sandbox to make this question testable in a phys
Physical understanding and reasoning depend on forming compact and generalizable representations of the world. While modern vision-language models can recognize and explain diverse physical events, they often lack explicit representations of the underlying mechanisms-such as object states, physical parameters, and governing dynamics-needed for reliably reasoning how the world evolves and responds
Recent video generation models are increasingly framed as world models. Many physical processes can unfold in more than one valid way. Therefore, a world model should reproduce not only a plausible trajectory, but also the distribution of possible behaviors under the same initial observation and action. We call this distribution-level requirement probabilistic alignment. However, existing evaluati
LLM agents increasingly rely on generated interaction data to learn how to interact with external environments. Agentic data generation must maintain consistency among environments, tasks, interactions, and success signals while producing experience that is useful rather than merely abundant. Existing work spans many agent domains, but domain-centered organization and heterogeneous evaluation ofte
Evolution Strategies (ES) have recently emerged as a memory-efficient post-training paradigm for LLM reasoning. However, the optimization behavior of ES remains understudied, making it hard to define its advantage scope compared to mainstream post-training paradigms (e.g., Group Relative Policy Optimization (GRPO)). By systematically investigating ES dynamics and mechanisms, this paper first ident
Conventional video editing primarily focuses on scene-level content, whereas live streaming places greater emphasis on the human subject. However, directly applying existing video-editing methods to human-centric live streaming remains challenging, as they may introduce facial-expression inconsistencies and typically depend on multiple offline inference steps, making them unsuitable for real-time
On-policy distillation (OPD), which leverages a pre-trained, specialized teacher model to provide dense supervisory signals, has achieved significant success in Large Language Models (LLMs) and has recently been adapted to flow matching models. However, this paradigm suffers from two major issues: First, training a separate, task-specific teacher for every new objective incurs high computational c
Explicit visual intermediates can help multimodal large language models (MLLMs) externalize spatial evidence and updated visual states, but their utility depends on whether an image editor can faithfully realize the required transformation. We introduce Aphanta, an automated task-discovery and closed-loop diagnostic framework for the MLLM -> image editor -> MLLM pipeline. Aphanta evaluates three c
Modern game development relies heavily on conventional graphics pipelines. High-quality visual content requires modeling, material authoring, animation, lighting, effects, and runtime optimization, making asset production expensive and extending the development cycle of game prototypes. Recently, video foundation models are beginning to change film and video production, but games differ from linea
Language model pretraining has become almost synonymous with prohibitive cost, placing it out of reach for much of the academic and open-source communities. Although strong open-source efforts already exist, including open-weight models and open-source training recipes, a cost-efficient, hardware-accessible, and open-source pretraining recipe has long been missing. Even at a small scale, training
While generative AI has significantly advanced video editing, existing methods primarily focus on single-shot or short video clips. Editing long videos with multiple instructions remains a formidable challenge. Naive chunking strategies, e.g., fixed-duration segmentation, often lead to entity fragmentation, severe editing hallucinations, and disrupted temporal continuity. To bridge this gap, we in
LLM judges are widely used to evaluate agentic tool-calling systems, yet their reliability on structured, dependency-driven workflows remains largely unexamined. We present AgentJudgeBench, the first benchmark to systematically study LLM-as-a-judge reliability for agentic tool-calling over workflow DAGs, as distinct from the broader LLM-as-a-judge task of open-ended text or preference evaluation.
Long autoregressive video generation faces a fundamental memory challenge: with a finite attention window, a model must decide which information from an ever-expanding history to retain. Existing methods organize memory temporally, preserving recent frames while compressing or discarding older ones. We instead propose RECAP-Forcing, organizing memory by appearance novelty. A long video is not mere
Large language models offer broad capabilities, but adapting them to evolving domains, tools, and requirements often entails repeated post-training. Autonomous systems automate parts of this process by proposing updates, training candidates, and using evaluation feedback to select subsequent proposals. As evidence accumulates, a central problem emerges: which past update evidence remains actionabl
Self-evolving language models have recently emerged as a promising path toward superintelligence, with the advantage of reducing the cost of human supervision. While considerable progress has been made in verifiable domains, self-evolution in unverifiable domains remains substantially less explored. We propose Judge co-adaptation from Zero data (J-Zero), a unified Challenger--Solver--Judge co-evol
Long-horizon agent runs generate experience that can improve both the current run and future work. Most self-improvement methods process this experience only after execution ends, so they cannot redirect the active run or immediately apply and validate lessons learned from it. We argue that self-improvement should instead be live, using emerging experience both to redirect the active run and to up
Scaling video generation to long durations reveals a critical bottleneck: current models lack robust long-term memory. This deficiency can be studied along two critical aspects: object permanence, the ability to precisely reproduce the appearance of objects upon re-entry; and memory capacity, the ability to process ultra-long context and use information from distant history. Robust long-term memor
Amazon Bedrock AgentCore Evaluations decouples agent evaluation from the framework you build on. As long as your agent emits OpenTelemetry telemetry, the service can score it, whether you use LangGraph, LlamaIndex, the OpenAI Agents SDK, Google ADK, the Claude Agent SDK, or Strands Agents. This post explains how the framework-agnostic contract works.
Health & Bioscience
Now you can get more intelligent speech-to-text transcription with Gemini 3.5 Transcribe.
In this post, you will learn how GoDaddy migrated from their legacy business intelligence (BI) tool to Amazon Quick. This was a two-year transformation that delivered results across every dimension of the business: 15,000 hours saved annually, 50% reduction in dashboard count, rendering times cut to under 5 seconds, and AI-powered self-service analytics now accessible to every employee.
Learn how Natera built an automated voice agent on Amazon Bedrock AgentCore that lets patients book mobile phlebotomy appointments through natural conversation. The post covers the dual-WebSocket bridge, event-driven latency masking, and progressive-trust authentication behind 100% tool-calling accuracy and sub-7-second latency.
The SageMaker Python SDK v3 redesigns script mode with unified ModelTrainer and ModelBuilder classes. This post walks through two end-to-end examples, a scikit-learn Random Forest and a multi-GPU Stable Diffusion 3.5 LoRA fine-tune, showing how SourceCode syncs your local code into any container at runtime so you can iterate without rebuilding Docker images.
The advanced side of supervised fine-tuning data prep. This second post in a two-part series covers evaluating data readiness with learning curves, selecting high-value data subsets, augmenting data with synthetic and distilled examples, and mixing data sources to prevent catastrophic forgetting.
Data preparation determines the ceiling of any supervised fine-tuning project. This first post in a two-part series covers the foundations of SFT data prep: quality checks, conversational (JSONL) formatting, reasoning and tool-calling schemas, and a representative train/evaluation split.
356 points, 361 comments
Learn how Amazon Bedrock AgentCore agents in one account can generate answers from an Amazon Bedrock knowledge base backed by Amazon Redshift Serverless in another account, without copying source data. This post covers the architecture, security boundary, and two orchestration models: a code-based Strands agent and a declarative AgentCore harness.
251 points, 171 comments
958 points, 482 comments
Editor's note: A product image was updated after initial publication. As AI takes on more complex work, business leaders face a new challenge: enabling rapid innovation using agents whil
657 points, 213 comments
257 points, 512 comments
423 points, 142 comments
ChatGPT for Teachers is expanding to 55 U.S. school systems, bringing secure AI tools, training, and support to over 100,000 more educators and staff.
OpenAI’s new report explores how students and educators use ChatGPT to make learning more continuous, with support that extends beyond the classroom.
9551 articles sourced historically · 100 per page