model releases

VFIG: Vectorizing Complex Figures in SVG with Vision-Language Models

arXiv:2603.24575v2 Announce Type: replace-cross Abstract: Scalable Vector Graphics (SVG) are essential for technical illustration and digital design, offering resolution independence and semantic edit

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

VibeWorlding: Can Multimodal Agents Construct 3D Open Worlds End-to-End?

DGX agent

arXiv:2608.15265v1 Announce Type: new Abstract: Constructing an interactive 3D open world from a user query is important. However, existing methods are primarily evaluated on idealized, simple queries

model-releasesarxiv-cs-ai

18 Aug 2026

VideoGAIA: A Benchmark for General AI Assistants on Agentic Video Understanding

arXiv:2608.14718v1 Announce Type: cross Abstract: Video understanding is a fundamental task for evaluating the capabilities of multimodal large language models (MLLMs). However, existing leading model

Vision Language Models Cannot Plan, but Can They Formalize?

arXiv:2509.21576v2 Announce Type: replace Abstract: The advancement of vision language models (VLMs) has empowered embodied agents to accomplish simple multimodal planning tasks, but not long-horizon

ViTaR: Visuo-Tactile Residual Adaptation for Foundation VLA Manipulation

model-releasesarxiv-cs-ro

arXiv:2608.15816v1 Announce Type: new Abstract: As Vision-Language-Action (VLA) models scale toward real-world deployment, contact-rich manipulation exposes a critical blind spot: these policies encod

Walk Before You Run: The Importance of Data Exploration for Data Analysis Agents

arXiv:2608.16045v1 Announce Type: cross Abstract: LLM-based data-analysis tools are increasingly used to help users analyze messy spreadsheets and workbooks, from answering questions over uploaded fil

WANDR: A Benchmark for Wide and Deep Research

arXiv:2608.14747v1 Announce Type: new Abstract: WANDR (Wide ANd Deep Research) is a benchmark of 500 realistic, challenging data-collection tasks for research agents. Each task requires a system to di

We just released TensorRT Model Connect in Public Preview. You can take a supported @huggingface model to end-to-end TensorRT inference in j…

model-releasesclem-delangue--x

We just released TensorRT Model Connect in Public Preview. You can take a supported @huggingface model to end-to-end TensorRT inference in just two commands. No intermediate ONNX export, and the resul

We still don’t know how people are really using AI

model-releasesmit-tech-review

AI companies like Anthropic and OpenAI regularly publish reports on how people are using products like Claude and ChatGPT, but they only release the data they want us to see, AI researchers say. “Ther

WeSCE: A Benchmark for Measuring Security Drift in LLM-Driven Code Editing

arXiv:2608.15092v1 Announce Type: cross Abstract: In this work, we introduce WeSCE, a benchmark for quantifying security drift in code editing under weak-security constraints, where tasks specify only

What can GPT 5.6 sol do that GPT 5.5 xhigh couldn’t ?? As always @GaryMarcus vindicated This isn’t a fucking issue about models getting “too…

model-releasesgary-marcus--x

What can GPT 5.6 sol do that GPT 5.5 xhigh couldn’t ?? As always @GaryMarcus vindicated This isn’t a fucking issue about models getting “too smart.” Quite the opposite. ChatGPT is delaying because the

What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models

arXiv:2608.16852v1 Announce Type: new Abstract: Regulatory compliance monitoring in deployed language models is increasingly implemented as a legal and audit control, checking model outputs against wr

What Does Context Compression Cost an Agent? Interaction Costs Unrevealed by Task-Completion Metrics

arXiv:2608.16370v1 Announce Type: new Abstract: Task completion is the standard metric for evaluating context compression, yet it is incomplete: compression can increase an agent's interaction cost by

What to Forget in Unlearning? Forget Set Curation for Language Models

arXiv:2608.14855v1 Announce Type: new Abstract: Machine unlearning aims to remove targeted data or behaviors from a trained model without retraining from scratch. Yet most evaluations assume that the

What's the best tool for offline Wikipedia RAG at the moment?

model-releasesr-localllama

Theoretical use case: I have a compressed version of wikipedia, and some function or tool fooRAG. Then I set my system prompt (or make a skill, or hook, etc.) to include 'You can search wikipedia with

When Agentic Executions Fail: Detecting and Localizing Runtime Faults from Telemetry

arXiv:2608.14680v1 Announce Type: new Abstract: Reliability in LLM-based agentic systems is a property of the whole execution (its tool calls, model calls, guardrails, and inter-agent messages), not o

When Do LLMs Apply the Wrong Law? Diagnosing LLM Failures in Temporal Legal Reasoning

arXiv:2608.14610v1 Announce Type: new Abstract: Legal reasoning tasks such as legal judgment prediction (LJP) require identifying the temporally correct version of the law governing a case -- a capabi

When do machine-learned exchange-correlation improvements inherit into density-functional tight binding?

arXiv:2608.14875v1 Announce Type: cross Abstract: Machine-learned exchange-correlation functionals correct band gaps at near-semilocal cost, while density-functional tight binding reaches the 10^3-10^

When Entropy Is Not Enough: Reclaiming Lost Semantics in LLM Output Length Prediction

arXiv:2608.15592v1 Announce Type: new Abstract: Efficient LLM serving is often bottlenecked by the need to pad sequences to a fixed maximum length, and this wastes compute and degrades throughput. Pre

When Less Is Enough: Context Selection and Prompting Strategies for Bengali News Headline Generation

arXiv:2608.15879v1 Announce Type: new Abstract: Large language models (LLMs) have shown strong performance in text generation tasks, yet their effectiveness on headline generation remains sensitive to

When Single-Dataset Conclusions Fail: A 45-Task Study of Threshold Tuning and Resampling for Imbalanced Classification

arXiv:2608.16147v1 Announce Type: new Abstract: Class-imbalance handling is routinely evaluated on a single benchmark dataset, and the resulting conclusions are reported as if they were properties of

When Stories Evolve: Benchmarking LLM Storytelling Across Agent Architectures in Open-Ended World Simulations

arXiv:2608.15654v1 Announce Type: cross Abstract: Large language models can write fluent stories, but open-ended storytelling requires more than local fluency. In evolving world simulations and AI-nat

When to Communicate: Belief Distributions and KL Divergence for Principled Gating in Multi-Agent RL

arXiv:2608.14559v1 Announce Type: new Abstract: Effective communication in multi-agent reinforcement learning requires agents to decide not only extit{what} to communicate, but when? Existing approach

When Tool-Backed Skill Retrieval Fails: Source-Style Collapse in Executable Capability Retrieval

arXiv:2608.16502v1 Announce Type: new Abstract: Large-scale agents increasingly rely on retrieval to access external capabilities. We study this retrieval gate in structured tools and APIs, a measurab

Whose Gold? Annotator-Pool Disagreement Is Large at the Item Level, and Hidden by Small Leaderboards

arXiv:2608.15980v1 Announce Type: new Abstract: Preference benchmarks are built by hiring annotators, and the identity of those annotators is treated as an implementation detail. We measure what that

Wiola 13M, a Gated Spiral Attention Architecture for Parameter Efficient Small Language Models

arXiv:2608.14604v1 Announce Type: cross Abstract: Small language models in the ten to one hundred million parameter range are attractive for on device inference, rapid experimentation, and controlled

😮 wow that’s huge 💰

model-releasesharrison-chase--x

😮 wow that’s huge 💰 🚀Today we launched LangSmith Tuned Evaluators, starting with Perceived Error. Tuned Evaluators run on production traces to catch undesirable agent behavior and attach feedback that

X^2Localizer: Cross-grained Alignment for Progressive Cross-view Video Geo-localization

arXiv:2608.16658v1 Announce Type: cross Abstract: Cross-view Video Geo-localization (CVG) aims to localize ground-view videos by retrieving their corresponding geo-tagged aerial images. However, CVG a

YOLO26-RD: An End-to-End Road Damage Detection Network With Learnable Contrast Enhancement and Edge-Guided Downsampling

model-releasesarxiv-cs-cv

arXiv:2608.15713v1 Announce Type: new Abstract: Automated pavement-distress detection is commonly framed as a small-object problem, motivating high-resolution P2/4 detection heads and lossless downsam

zLend: A Dual-Scope Cash-Flow Reconstruction Framework for On-Chain Credit Underwriting

arXiv:2608.16856v1 Announce Type: cross Abstract: Decentralized lending lacks a credit bureau: a borrower's capacity to repay must be inferred entirely from public on-chain activity, without income ve

100$ worth of gpu runs qwen 3.8 27b at 7.39 t/s

model-releasesr-localllama

Qwen 27b Q3_K_M 2x rx 580 8gb (~50 each in my country, edge cases 60 per gpu) gives us 16gb vram We used it on an old already existing ddr3 motherboard with 2 gpu slots(you can buy it ror around 200$

1/2 the turns and 25% of the tokens. Task/Token is something to watch. Is Grok 4.6 inside Grok Bot? Has this been revealed yet? Have I misse…

model-releaseselon-musk--x

1/2 the turns and 25% of the tokens. Task/Token is something to watch. Is Grok 4.6 inside Grok Bot? Has this been revealed yet? Have I missed it? I’ve been wondering what magic allows grok bot to just

A Calibrated Test of Internal Action Maps: State Signals Without Global Affine Closure

arXiv:2608.13626v1 Announce Type: new Abstract: A hidden state signal can be decodable or causally usable without supporting a reusable action map. We test whether action maps fitted without a source

A Four-Axis Trustworthiness Benchmark for LLM-as-Judge in Principle-Based Regulation

arXiv:2608.14329v1 Announce Type: cross Abstract: Principle-based regulation, with evaluative standards such as 'fair, clear, and not misleading' or 'deliver good outcomes', cannot be reduced to binar

A Graph-Based Reinforcement Learning Framework for Structured Drift Diagnosis and Recovery in Autonomous LLM Agents

arXiv:2608.14109v1 Announce Type: new Abstract: Autonomous LLM agents are increasingly deployed in complex real-world workflows, yet they remain vulnerable to runtime behavioral drift, a silent deviat

A Pathway to General-Purpose Scientific AI: Multimodal Comprehension of Scientific Images

arXiv:2608.14075v1 Announce Type: new Abstract: Scientific figures and tables encode essential experimental evidence, yet remain difficult for digital libraries and multimodal AI systems to retrieve a

A Unified Assessment of the Poverty of the Stimulus Argument for Neural Language Models

arXiv:2602.09992v2 Announce Type: replace-cross Abstract: Several recent contributions have evaluated the Poverty of the Stimulus Hypothesis (PoSH) using Artificial Neural Networks (ANNs). The results

Act2Intention: A Benchmark For Developing Active Mobile Agents Through Inferring User Intention from GUI Actions

arXiv:2608.14132v1 Announce Type: cross Abstract: Mobile GUI Agents powered by multimodal large language models (MLLMs) show promise in human-computer intelligence. However, current research primarily

Adaptive Protection for Evolutionary Feature Construction in Symbolic Regression with Application to Credit Classification

arXiv:2608.14209v1 Announce Type: new Abstract: Evolutionary feature construction has shown strong promise in symbolic regression by automatically discovering informative transformations of input feat

After pushing 1M+ tokens through Qwen 3.8 27B, here is my optimal llama.cpp config for 16GB VRAM (73k Context, Agentic Coding)

model-releasesr-localllama

Following up on my previous post about my budget server setup (Intel N100 + RTX 5060 Ti 16GB), a few of you asked for a deeper dive into my actual inference config and real-world agentic performance.

Agentic Transaction: Towards ACID-Compliant Agent Systems

arXiv:2608.13900v1 Announce Type: cross Abstract: Large language model (LLM) agents are evolving from conversational assistants into autonomous systems that execute long-horizon tasks through reasonin

AgentRewind: Recoverable Execution for Long-Horizon LLM Agents

arXiv:2608.14380v1 Announce Type: new Abstract: Many real-world tasks require LLM agents to interact with their environments over long execution horizons. Errors that occur early in execution may prop

AI-Assisted Discovery and Construction of a Counterexample to the Convergence of Three-Block ADMM with the Identity Matrix as its Third Constraint Block

arXiv:2608.14396v1 Announce Type: cross Abstract: The alternating direction method of multipliers (ADMM), as a landmark algorithm, has attracted tremendous research attention and extensive practical a

AI Research Preference Models

arXiv:2608.13940v1 Announce Type: new Abstract: AI research agents (AIRA) can now propose, implement, and evaluate their own machine learning experiments, but progress on frontier tasks is throttled b

Alibaba's open-weight AI models accumulated 3 billion global downloads in the past 6 months — more than Meta Llama, more than Google Gemma, …

model-releasesclem-delangue--x

Alibaba's open-weight AI models accumulated 3 billion global downloads in the past 6 months — more than Meta Llama, more than Google Gemma, more than every Chinese domestic competitor combined. The mo

AnchorBench: A Multi-Pathway Benchmark for the Anchoring Effect in LLMs

arXiv:2608.14320v1 Announce Type: new Abstract: The anchoring effect is a cognitive bias in which an initial reference value shifts a later judgment toward itself. This effect is well established in h

Anthropic explains how Claude’s invisible text watermarks will work

model-releasesthe-verge-ai

Anthropic has clarified how it's planning to apply invisible watermarks to Claude-generated text in order to comply with Europe's AI transparency rules. On Friday, Anthropic announced that Claude's te

Anthropic's text watermark alters word probabilities to embed a fingerprint, which could degrade Claude's writing, despite its claim of no impact on quality (John Gruber/Daring Fireball)

John Gruber / Daring Fireball: Anthropic's text watermark alters word probabilities to embed a fingerprint, which could degrade Claude's writing, despite its claim of no impact on quality — When I wro

model-releasestechmeme