model releases

Gemma 4 Chat Template now has preserve thinking

model-releasesr-localllama

Google added an empty thinking token to the Gemma 4 chat template, which stabilizes model output by suppressing 'ghost' thought channels that may appear even when thinking is deactivated. This update

GlucoFM-Bench: Benchmarking Time-Series Foundation Models for Blood Glucose Forecasting

DGX agent

arXiv:2606.06881v1 Announce Type: new Abstract: Blood glucose forecasting models are foundational for modern diabetes management systems, as reliable short-term predictions can enable proactive interv

model-releasesarxiv-cs-lg

8 Jun 2026

Google and Nvidia are helping Apple with Apple Foundation Model Cloud Pro, which Apple says is comparable to Gemini frontier models and runs on Nvidia GPUs (Kif Leswing/CNBC)

Kif Leswing / CNBC: Google and Nvidia are helping Apple with Apple Foundation Model Cloud Pro, which Apple says is comparable to Gemini frontier models and runs on Nvidia GPUs — - Apple showed off Sir

model-releasestechmeme

Google upgrades NotebookLM, which now runs on Gemini 3.5 and Antigravity, to deliver new agentic capabilities and more advanced reasoning for AI Ultra users (Ivan Mehta/TechCrunch)

Ivan Mehta / TechCrunch: Google upgrades NotebookLM, which now runs on Gemini 3.5 and Antigravity, to deliver new agentic capabilities and more advanced reasoning for AI Ultra users — Google on Monday

model-releasestechmeme

Great tips. In practice, this is how it roughly looks to run agents autonomously for hours or days. /goal or /loop to keep it going. Verific…

Great tips. In practice, this is how it roughly looks to run agents autonomously for hours or days. /goal or /loop to keep it going. Verification is crucial here. Seeing a number of benchmarks showing

model-releasesdair-ai--x

GS-KAN: Parameter-Efficient Kolmogorov-Arnold Networks via Sprecher-Type Shared Basis Functions

arXiv:2512.09084v3 Announce Type: replace Abstract: The Kolmogorov-Arnold representation theorem offers a theoretical alternative to Multi-Layer Perceptrons (MLPs) by placing learnable univariate func

GuideCAD: A Lightweight Multimodal Framework for 3D CAD Model Generation via Prefix Embedding

model-releasesarxiv-cs-cv

arXiv:2606.07024v1 Announce Type: new Abstract: Multi-modal approaches used for 3D CAD generation require substantial computational resources, necessitating efficient training. To address this, we pro

Hierarchical Certified Semantic Commitment for Byzantine-Resilient LLM-Agent Collaboration

arXiv:2606.07316v1 Announce Type: cross Abstract: Byzantine collaboration among large-language-model agents requires a finality-control primitive: given delivered stochastic, structured natural-langua

Hierarchical Semantic-Constrained Heterogeneous Graph for Audio-Visual Event Localization

arXiv:2606.07033v1 Announce Type: new Abstract: Open-vocabulary audio-visual event localization (OV-AVEL) jointly models audio-visual cues to recognize and temporally localize events, including catego

HKJudge: A Legal Discourse-Annotated Corpus for Interpreting What Courts Find, How They Reason, and What They Rule

arXiv:2606.06679v1 Announce Type: cross Abstract: Court judgments are central to legal practice and jurisprudence, yet discourse analysis of Hong Kong judgments has received limited attention, owing l

Improving Cross-Lingual Factual Recall via Consistency-Driven Reinforcement Learning

arXiv:2606.06586v1 Announce Type: new Abstract: Large language models (LLMs) trained predominantly on English data encode substantial world knowledge, yet often fail to express it reliably in other la

Inheritance Between Feedforward and Convolutional Networks via Model Projection

arXiv:2602.06245v2 Announce Type: replace-cross Abstract: Neural-network techniques are often transferred across architecture families by analogy, but such transfer is valid only when the assumptions

Interesting model here 35b a3b trained for agentic use It gets 60.7 on Terminal Bench2 qwen 3.6 27b gets 59.3 Essentially the same Going to …

model-releasesclem-delangue--x

Interesting model here 35b a3b trained for agentic use It gets 60.7 on Terminal Bench2 qwen 3.6 27b gets 59.3 Essentially the same Going to have to try it out https://huggingface.co/nex-agi/Nex-N2-min

Introducing 'Signals of AI' - DeepAgent powered by - @nebiustf for Inference - @NVIDIAAI Nemotron 3 Ultra - @tavilyai for Web Search - @Lang…

model-releasesharrison-chase--x

Introducing 'Signals of AI' - DeepAgent powered by - @nebiustf for Inference - @NVIDIAAI Nemotron 3 Ultra - @tavilyai for Web Search - @LangChain for the core deepagent architecture! An ideal engine f

Introducing the OpenAI Economic Research Exchange

The OpenAI Economic Research Exchange is an initiative by OpenAI designed to facilitate collaboration between economists and AI researchers to study the economic impacts and implications of artificial

model-releasesopenai

It's a big AI day for Apple at WWDC26! A few humble name proposals for the new overhaul: - Apple Intelligence 2: Siri’s Revenge - Apple Inte…

model-releasesrowan-cheung--x

It's a big AI day for Apple at WWDC26! A few humble name proposals for the new overhaul: - Apple Intelligence 2: Siri’s Revenge - Apple Intelligence: 2 Fast 2 Furious - Apple Intelligence II (ft. Gemi

It's a TRAP! Task-Redirecting Agent Persuasion Benchmark for Web Agents

arXiv:2512.23128v2 Announce Type: replace-cross Abstract: Web-based agents powered by large language models are increasingly used for tasks such as email management or professional networking. Their r

It’s safe to close your laptop now: Hosting coding agents on Amazon Bedrock AgentCore

model-releasesaws-ml-blog

Amazon Bedrock AgentCore Runtime gives each agent session its own isolated microVM with a persistent workspace, secure tool access through Gateway, and built-in observability—so you can run Claude Cod

Last time around Apple released a lot of information about how their AI version of Siri worked between local and cloud models, not so much t…

model-releasesethan-mollick--x

Last time around Apple released a lot of information about how their AI version of Siri worked between local and cloud models, not so much this time It is nice to have a Gemma-like model on device, bu

Lean4Agent: Formal Modeling and Verification for Agent Workflow and Trajectory

arXiv:2606.06523v1 Announce Type: new Abstract: Equipping Large Language Models (LLMs) to execute reliable multi-step workflows has become a central challenge in artificial intelligence. Despite recen

Learn to Match: Two-Sided Matching with Temporally Extended Feedback

arXiv:2606.06744v1 Announce Type: new Abstract: Two-sided matching markets often involve information that unfolds over time through interviews, repeated interaction, learning, and separation. Existing

Learning Perspectivist Social Meaning via Demographic-Conditioned Fusion Embeddings

arXiv:2606.07123v1 Announce Type: new Abstract: Social meaning in language is inherently perspectival, varying across annotator backgrounds, demographics, and ideological positions. However, most NLP

LiQSS: Post-Transformer Linear Quantum-Inspired State-Space Tensor Networks for Real-Time 6G

arXiv:2601.12375v3 Announce Type: replace-cross Abstract: Proactive and agentic control in Sixth-Generation (6G) Open Radio Access Networks (O-RAN) requires control-grade prediction under stringent Ne

LLM-Guided Evolution for Medical Decision Pipelines

arXiv:2606.07342v1 Announce Type: new Abstract: Adapting large language models (LLMs) to clinical workflows often requires costly fine-tuning or manual prompt and pipeline engineering. We study LLM-gu

M^3Exam: Benchmarking Multimodal Memory for Realistic User-Agent Interactions

arXiv:2606.07402v1 Announce Type: new Abstract: Language agents are increasingly deployed over accumulating multimodal information, yet existing benchmarks assume a human-human form with sparse visual

MacArena: Benchmarking Computer Use Agents on an Online macOS Environment

arXiv:2606.06560v1 Announce Type: cross Abstract: Computer-use agents (CUAs) operate graphical user interfaces (GUIs) through vision and control primitives, and their capabilities have advanced rapidl

MACD: Model-Aware Contrastive Decoding via Counterfactual Data

arXiv:2602.01740v3 Announce Type: replace Abstract: Video language models (Video-LLMs) are prone to hallucinations, generating plausible but ungrounded content when visual evidence is weak, ambiguous,

Machine Learning for Electron-Scale Turbulence Modeling in W7-X

arXiv:2511.04567v2 Announce Type: replace-cross Abstract: Constructing reduced models for turbulent transport is essential for accelerating profile predictions and enabling many-query tasks such as pa

MADE: Beyond Scoring via a Multilingual Agentic Diagnosing Engine for Fine-Grained Evaluation Insights

arXiv:2606.07020v1 Announce Type: new Abstract: Multilingual and multicultural benchmarks now cover dozens of languages and model families, but the resulting score landscapes remain metric-rich and in

Making the Most of Limited Data: Score-Aware Training for Text-to-Music Generation

arXiv:2606.07387v1 Announce Type: new Abstract: State-of-the-art text-to-music generation systems rely on massive proprietary datasets and industrial-scale compute, making it impossible to disentangle

MatterDoor: Sampling Zero-shot Spatio-semantic Priors using Generative Models

arXiv:2510.11014v2 Announce Type: replace-cross Abstract: Autonomous robots often view rooms only partially, through a doorway, where the walls and scene structure hide the geometry and task-relevant

MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism

arXiv:2606.07512v1 Announce Type: cross Abstract: Current Vision-Language Models struggle with hours-long videos because processing full-length visual sequences induces prohibitive token explosion and

Meta puts $115M toward a 'workforce academy', a free 5-week program to train Americans to build data centers, offering job guarantees at Meta construction sites (Wall Street Journal)

Wall Street Journal: Meta puts $115M toward a “workforce academy”, a free 5-week program to train Americans to build data centers, offering job guarantees at Meta construction sites — Five-week progra

model-releasestechmeme

MMAE: A Massive Multitask Audio Editing Benchmark

arXiv:2606.07229v1 Announce Type: cross Abstract: We introduce MMAE, a Massive Multitask Audio Editing benchmark, serving as the first comprehensive evaluation testbed designed for general-purpose ins

MMBU: A Massive Multi-modal Biomedical Understanding Benchmark to Probe the Perception Capabilities of Vision-Language Models

arXiv:2606.06696v1 Announce Type: cross Abstract: Vision and language models (VLMs) hold immense promise to transform biomedical imaging workflows, from detecting lesions in chest X-rays to profiling

mmPISA-bench: Do LLMs Reason Equally Well Across 43 Languages?

arXiv:2606.07069v1 Announce Type: new Abstract: We introduce mmPISA-bench, a compact high-quality multilingual reasoning benchmark derived from the OECD Programme for International Student Assessment

Model Recycling Framework for Multi-Source Data-Free Supervised Transfer Learning

arXiv:2508.02039v2 Announce Type: replace Abstract: Increasing concerns for data privacy and other difficulties associated with retrieving source data for model training have created the need for sour

Modeling Nonlinear Feature Interactions with Product-Unit Residual Networks

arXiv:2606.06861v1 Announce Type: cross Abstract: Understanding nonlinear feature interactions is crucial in science and engineering, yet standard multilayer perceptrons (MLPs) often capture such inte

Modernizing Healthcare: How Alcidion achieved greater stability and performance with AlloyDB

model-releasesgoogle-cloud-ai

In clinical informatics, every second counts. For Alcidion, a global leader in smart health solutions, the mission is simple but critical: use technology to reduce cognitive load for clinicians and pr

MotionEnhancer: Leveraging Video Diffusion for Motion-Enhanced Vision-Language Models

arXiv:2606.06853v1 Announce Type: cross Abstract: The new era has witnessed a remarkable capability to extend Vision-Language Models (VLMs) for tackling tasks of video understanding. While current VLM

mtmd : add video input support by ngxson · Pull Request #24269 · ggml-org/llama.cpp

model-releasesr-localllama

PR #24269 added native video input to llama.cpp's multimodal (mtmd) system, merging on June 8, 2026. The implementation uses FFmpeg as a subprocess to decode video frames and expands a single video ma

Multi-Objective Preference Optimization: Improving Human Alignment of Generative Models

arXiv:2505.10892v2 Announce Type: replace Abstract: Post-training LLMs with RLHF and preference optimization methods (e.g., DPO, IPO) has greatly improved alignment, yet these approaches assume a sing

Never Seen Before: Benchmarking Genuine Zero-Shot Composed Image Retrieval with Consistent Video-Sourced Datasets

arXiv:2606.07032v1 Announce Type: cross Abstract: Zero-Shot Composed Image Retrieval (ZS-CIR) aims to retrieve a target image based on a query composed of a reference image and a relative caption with

NotebookLM’s Gemini 3.5 upgrade adds a cloud computer and help finding sources

model-releasesthe-verge-ai

Google is rolling out 'across the board' updates to NotebookLM. The AI-powered note-taking app now uses Google's upgraded Gemini 3.5 model, which will allow it to respond with 'more accurate and relia

NTILC: Neural Tool Invocation via Learned Compression

arXiv:2606.06566v1 Announce Type: cross Abstract: Agentic tool-calling language models depend on large registries of callable APIs, functions, and local actions. Placing full tool specifications direc

Oh my God! @METR_Evals’s coding benchmarks are saturated! 🤯 Mythos broke the METR graph 🤯 4 weeks later, out comes a new coding task, this…

model-releasesgary-marcus--x

Oh my God! @METR_Evals’s coding benchmarks are saturated! 🤯 Mythos broke the METR graph 🤯 4 weeks later, out comes a new coding task, this time from @cognition: “FrontierCode Diamond remains unsaturat

On the Effect of Neural Field Reparameterization for 4DVAR

arXiv:2509.21751v2 Announce Type: replace Abstract: Four-dimensional variational data assimilation (4DVAR) is a cornerstone of numerical weather prediction, yet it remains computationally intensive an

On the Geometry of On-Policy Distillation

arXiv:2606.07082v1 Announce Type: cross Abstract: On-policy distillation (OPD) is increasingly used to improve large language model reasoning, but its training dynamics remain poorly understood. We ch