Large language model news
The open-weight and frontier model landscape changes weekly. This page follows new large language models from Meta, Mistral, Alibaba, DeepSeek and smaller labs, together with benchmark results, licence changes and the research behind each release.
Latest stories
Researchers say LLMs can be good at natively reading and writing neural network weights; GPT-6 Astra edited Qwen3.5-0.8B's weights to make it better at math (Rikiya Takehi)
Rikiya Takehi : Researchers say LLMs can be good at natively reading and writing neural network weights; GPT-6 Astra edited Qwen3.5-0.8B's weights to make it better at math — TL;DR … If you were asked to handcraft a neural network that plays Go, what would you do? If you were handed raw neural network weights …
Techmeme
Deep Principle Releases Hydrogen-1.0: A 753B Open-Weight Model for Scientific Discovery
Deep Principle has released Hydrogen-1.0, an open-weight scientific model for chemistry, materials science and life sciences. It is built by post-training Z.ai’s GLM-5.3 on domain data, laboratory records and agent discovery trajectories. For developers, it offers open weights that report a FrontierScience-Research win over GPT-6-Astra. The weights ship under the MIT License. TL;DR What is […] The post Deep Principle Releases Hydrogen-1.0: A 753B Open-Weight Model for Scientific Discovery appeared first on MarkTechPost .
MarkTechPost
DeepSeek releases a "creator mode" that allows DeepSeek Harness users to quickly create custom plugins, like a 3D model renderer and a video editing workbench (Debby Wu/Bloomberg)
Debby Wu / Bloomberg : DeepSeek releases a “creator mode” that allows DeepSeek Harness users to quickly create custom plugins, like a 3D model renderer and a video editing workbench — DeepSeek has unveiled a new option to quickly create custom plugins that help users to tackle a range of complex tasks …
Techmeme
MarkTechPost
AI Global WireMeet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Run CUDA
qwenGPU compiles a 4-lane SIMT GPU into Qwen3.5-2B as 45 untrained layers. The model writes every step of real CUDA launches itself, using Gated DeltaNet state as registers and memory. The post Meet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Run CUDA appeared first on MarkTechPost .
MarkTechPost
arXiv cs.AI
AI Global WireTowards Portable Specification of the Agent Harness
arXiv:2610.40917v1 Announce Type: new Abstract: Large language models act only through the infrastructure that surrounds them. That infrastructure, the harness, determines what a model can accomplish and how it fails, yet it has no settled definition. To address this gap, we decompose the harness into eight functional components that can be specified and compared across systems, and we propose a portable specification framework that records which model ran in which harness. The record is needed because different harness configurations can yield several-fold differences in reported scores for the same model, and because much of an agent's attack surface lies in its harness. Existing model car
arXiv cs.AI
arXiv cs.AI
AI Global WireP-ContextBench: From Personal Fact Recall to Personalized Agent Behavior
arXiv:2610.40959v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for personal tasks, yet evaluating their safe use of personal context remains difficult because such contexts are private, fragmented, dynamic, and socially situated. We introduce P-ContextBench, a realistic, verifiable, and scalable benchmark for evaluating personal-context understanding. P-ContextBench constructs population-structured, life-story-grounded digital humans from demographic seeds reflecting the U.S. population. Rather than using explicit, static profiles, P-ContextBench embeds traits, routines, goals, and constraints as implicit evidence across fragmented, evolving synthetic mobi
arXiv cs.AI
arXiv cs.AI
AI Global WireConsensus Within, Competition Between: Decentralized LLM Teams for Perimeter Defense
arXiv:2610.41044v1 Announce Type: new Abstract: The use of large language models (LLMs) to coordinate multi-agent teams raises a fundamental architectural question: whether task allocation should be centralized in a single LLM controller or distributed across autonomous LLM agents, particularly when communication is unreliable. In a deliberately simplified perimeter-defense simulation against LLM-controlled intruders, we compared a central commander, an auction-based consensus team and independent agents under five network conditions. Breach rates were 15% for the consensus team, 32% for the commander and 29% for independent agents, and of the different network conditions, only limited radio
arXiv cs.AI
arXiv cs.AI
AI Global WireCalibration of Large Model Quantization based on HiFloat4 Format
arXiv:2610.41112v1 Announce Type: new Abstract: Quantization of Large Language Models (LLMs) has be- coming increasingly important for reducing the computa- tional costs and resource requirements for practical deploy- ment.Growinginterestinthisareahasmotivatedthecommu- nity to move beyond 8-bit quantization towards 4-bit preci- sion, leading to the development of customized 4-bit formats such as MXFP4 and NVFP4. Recently, HiFloat4 (HiF4) has emerged as a competitive alternative which adopts a Block Floating-Point(BFP)formatwithathree-levelscalinghierar- chy, thereby achieving a significantly larger dynamic range. While quantization can substantially reduce memory usage and accelerate inferen
arXiv cs.AI
arXiv cs.AI
AI Global WireAgentic World Modeling for Agent Test-time Scaling without Environment Replica
arXiv:2610.40927v1 Announce Type: new Abstract: Test-time scaling (TTS) improves large language model agents by sampling and selecting among multiple candidate trajectories, but common realizations assume that these trajectories can be executed in parallel. This assumption breaks in realistic deployments, where enterprise workspaces and cloud services are shared and persistent, costly or infeasible to replicate, and actions may introduce irreversible side effects or cross-trajectory interference. Existing generative world models avoid direct execution, but typically compress environment state into a single prompt, limiting their ability to reliably simulate transitions under partial observab
arXiv cs.AI
arXiv cs.AI
AI Global WireIncongruence-Aware Evaluation of Multimodal Speech Emotion Recognition in Audio Large Language Models
arXiv:2610.40883v1 Announce Type: new Abstract: Voice-interactive large language models process both what a speaker says and how they say it, yet the two channels do not always agree. We ask whether that disagreement degrades speech emotion recognition. Mapping text and audio into a shared valence-arousal-dominance space, we convert the distance between them into a continuous, threshold-free incongruence score with a two-component Gaussian mixture, and validate the score against human judgement. We evaluate three audiocapable models on acted, conversational and spontaneous speech (eNTERFACE'05, MELD and CMU-MOSEI). With the model held constant across modalities, incongruence raises the odds
arXiv cs.AI
arXiv cs.AI
AI Global WireBABA-Bench: Evaluating Long-Horizon Patient-Advocacy Agents
arXiv:2610.41076v1 Announce Type: new Abstract: An LLM agent that helps a patient get care must do two things. It must reach the correct outcome, and it must follow the necessary procedure. For example, the agent must get the patient's permission before it shares the patient's records. We introduce BABA-Bench, a benchmark of 47 interactive worlds. We built each world from a de-identified, real patient advocacy case that expert advocates reviewed. In each world, the agent must find hidden information and coordinate insurers, clinics, suppliers, pharmacies, and community organizations. The engine checks each action at the time the agent does it. It checks if the agent has permission, if the ag
arXiv cs.AI
arXiv cs.AI
AI Global WireChatGPT as a Surveillance Assistant? An Exploratory Case Study of AI-Assisted Surveillance Risks
arXiv:2610.40890v1 Announce Type: new Abstract: Large language models (LLMs) can generate detailed guidance, functional code, and persuasive text at scale. As one of the most widely used consumer language model products as of 2026, ChatGPT gives broad public access to these capabilities, motivating concerns about its potential use in surveillance-related tasks. We present an exploratory case study of what potentially harmful surveillance assistance looks like in ChatGPT's public web interface. We contribute a taxonomy of seven surveillance risk categories based on prior LLM risk research and reports of deployed surveillance systems, instantiated as 110 prompts, along with a manually reviewed
arXiv cs.AI
arXiv cs.AI
AI Global WireThe psychophysics of affect in large language models
arXiv:2610.40985v1 Announce Type: new Abstract: Good decisions often require considering how outcomes will feel. Yet large language models (LLMs) increasingly enter decision workflows, despite little being known about how they map outcomes to feelings or use those mappings in choice. We tested how LLMs anticipated gambling or stock-trading outcomes would feel and whether they incorporated these ratings into choice. When outcomes were rated in isolation, ratings exhibited diminishing sensitivity to magnitude, resembling human value functions; but models rated gains more intensely than equivalent losses, an inversion of canonical loss aversion we call silicon gain amplification. When outcomes
arXiv cs.AI
arXiv cs.AI
AI Global WireCIRCE Divines LLM Task Complexity from Constraint Structure
arXiv:2610.41089v1 Announce Type: new Abstract: Estimating whether a problem is too complex for a large language model (LLM) before the model attempts it can guide model routing, how to allocate test-time compute, and suggest the need for new tools to aid the agent. While existing estimators rely on the problem text, LLM-annotated task difficulty labels, or signals obtained after a model attempts to solve the problem, such as its hidden states, none explicitly leverage the constraint structure inherent in the problem: the variables to assign and how the constraints over them interact. We introduce CIRCE, which extracts a problem's structure by using a cheap LLM to construct a Constraint Inte
arXiv cs.AI
arXiv cs.AI
AI Global WireOpening the Box Changes the Cat: LLMs Detect Chain-of-Thought Monitoring - What now?
arXiv:2610.40946v1 Announce Type: new Abstract: Chain-of-thought (CoT) reasoning allows large language models (LLMs) to generate intermediate reasoning traces in natural language while solving complex tasks. These traces can support safety monitoring by exposing potentially harmful plans before actions are executed. In this work, we investigate the conditions under which LLMs infer the presence of CoT monitoring from direct and indirect environmental cues and examine how this inference affects their subsequent behavior. Our observations suggest that oversight is an interactive process: while a monitor inspects a model's reasoning, traces left by monitoring interventions can allow the model t
arXiv cs.AI
arXiv cs.AI
AI Global WireConfident Fools or Careful Searchers? A Behavioural Study of When LLMs Search Beyond Their Static Internal Memory
arXiv:2610.41124v1 Announce Type: new Abstract: Agentic LLM assistants must repeatedly decide whether to answer from memory or to call a tool such as web search. Unnecessary searches cost latency and money; missing ones yield confident but unsupported answers. Prior work mostly measures verbalised confidence or trains new adaptive-retrieval methods, and rarely audits the search decisions that deployed products actually make. We introduce a benchmark of 400 multiple-choice questions that combines widely known inside-knowledge (IK) questions, which should be answered directly, with outside-knowledge (OK) questions about facts from news published after the models' training cut-offs (to our know
arXiv cs.AI
Iluvatar CoreX Tiangai 300 tops Nvidia H200 in select AI inference tests, with long-context and MoE gains
Chinese AI chipmaker Iluvatar CoreX has released benchmarks showing its next-generation Tiangai 300 general-purpose GPU outperforming Nvidia's H200 in selected large language model (LLM) inference, long-context processing and multimodal generation workloads.
DIGITIMES
TechNews (TW)
AI Global WireDeepSeek 發表程式設計代理 Super-Agent 不需指令即可主動修錯，正面挑戰 Claude Code
中國人工智慧新創 DeepSeek 旗下 Harness 團隊 10 月 31 日微信公眾號發表開源程式設計代 […]
TechNews (TW)
Inteligencia artificial (ES)
AI Global WireLa inteligencia artificial creada en Galicia que ha llamado la atención de la ONU: está entre las 20 mejores del mundo
Una plataforma desarrollada en una universidad gallega ha sido seleccionada entre 173 candidaturas de 50 países por su capacidad para hacer más ...
Inteligencia artificial (ES)
DeepSeek unveils AI ‘super agent’ in coding race with OpenAI, Anthropic
Chinese artificial intelligence start-up DeepSeek has previewed a plan for a more autonomous AI agent that can handle multiple tasks with minimal human intervention, as it steps up competition with US rivals OpenAI and Anthropic in developing coding assistants. The proposed system, dubbed super agent, aimed to reduce the need for humans to coordinate multiple AI agents and monitor their progress, addressing a key bottleneck in managing today’s increasingly complex workflows, according to a...
SCMP Tech
MarkTechPost
AI Global WireDeepSeek Open-Sources dsh-superagent, a Router-Worker Coding Agent
DeepSeek released dsh-superagent, an MIT-licensed research prototype that turns GitHub and Feishu requests into parallel worker sessions. In DeepSeek’s own simulation, p95 task completion fell to 15.7 minutes from 29.5 minutes. The post DeepSeek Open-Sources dsh-superagent, a Router-Worker Coding Agent appeared first on MarkTechPost .
MarkTechPost
DeepSeek builds ‘super-agent’ for autonomous coding
The system can try to fix bugs reported in group chats even without being directly tagged.
Tech in Asia
Kunstmatige intelligentie (NL)
AI Global WireDeepSeek daagt Anthropic uit in AI-coderingsstrijd
DeepSeek daagt Anthropic uit met autonome AI-coderingsagent De strijd om de heerschappij in kunstmatige intelligentie bereikt een nieuw hoogtepunt met de lancering van een prototype voor een geavancee ...
Kunstmatige intelligentie (NL)
Mistral AI publicly releases Large 4, targets cybersecurity users
Mistral said the released version was an upgrade from the one unveiled three weeks ago, following testing with governments and corporate users, which helped improve its performance.
Economic Times Tech
Apple, Alibaba to roll out Qwen-powered AI features in China
Testing of the Qianwen-powered features begins November 4 in the latest iOS developer beta.
Tech in Asia
