model releases

SWAN: Semantic Watermarking with Abstract Meaning Representation

arXiv:2605.04305v1 Announce Type: new Abstract: We introduce SWAN (Semantic Watermarking with Abstract Meaning Representation), a novel framework that embeds watermark signatures into the semantic str

Model Releases

Syntax- and Compilation-Preserving Evasion of LLM Vulnerability Detectors

DGX agent

arXiv:2602.00305v2 Announce Type: replace-cross Abstract: LLM-based vulnerability detectors are increasingly deployed in CI/CD security gating, yet their resilience to evasion under syntax- and compil

model-releasesarxiv-cs-lg

7 May 2026

TabEmbed: Benchmarking and Learning Generalist Embeddings for Tabular Understanding

arXiv:2605.04962v1 Announce Type: new Abstract: Foundation models have established unified representations for natural language processing, yet this paradigm remains largely unexplored for tabular dat

TCM-Serve: Modality-aware Scheduling for Multimodal Large Language Model Inference

arXiv:2603.26498v2 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) power platforms like ChatGPT, Gemini, and Copilot, enabling richer interactions with text, images, an

Telegraph English: Semantic Prompt Compression via Structured Symbolic Rewriting

arXiv:2605.04426v1 Announce Type: new Abstract: We introduce Telegraph English (TE), a prompt-compression protocol that rewrites natural language into a symbol-rich, formally-structured dialect. Where

Terminus-4B: Can a Smaller Model Replace Frontier LLMs at Agentic Execution Tasks?

arXiv:2605.03195v1 Announce Type: new Abstract: Modern coding agents increasingly delegate specialized subtasks to subagents, which are smaller, focused agentic loops that handle narrow responsibiliti

Testing ads in ChatGPT

OpenAI announced a test of advertisements within ChatGPT, marking the company's exploration of ad-supported monetization models alongside its existing subscription offerings. The initiative aims to ba

model-releasesopenai

The Chrome extension expands what Codex can do for coding and work. From debugging browser flows to checking dashboards, conducting research…

The Chrome extension expands what Codex can do for coding and work. From debugging browser flows to checking dashboards, conducting research, or updating CRMs, Codex can take on more of the tasks that

model-releasesopenai--x

The most female-led product org in tech right now: Chief Product Officer: Ami Vora Claude Code/Cowork Head of Product: Cat Wu Claude Code/Co…

model-releasesjeremy-howard--x

The most female-led product org in tech right now: Chief Product Officer: Ami Vora Claude Code/Cowork Head of Product: Cat Wu Claude Code/Cowork Head of Eng: Fiona Fung Claude Platform Head of Product

The Predictive-Causal Gap: An Impossibility Theorem and Large-Scale Neural Evidence

arXiv:2605.05029v1 Announce Type: new Abstract: We report a systematic failure mode in predictive representation learning. Across 2695 neural network configurations trained to predict linear-Gaussian

The Shape of Beliefs: Geometry, Dynamics, and Interventions along Representation Manifolds of Language Models' Posteriors

arXiv:2602.02315v2 Announce Type: replace Abstract: Large language models (LLMs) form implicit beliefs (posteriors over latent variables) from prompts, but we lack a mechanistic account of how these b

Tightly-Coupled Estimation and Guidance for Robust Low-Thrust Rendezvous via Adaptive Homotopy

model-releasesarxiv-cs-ro

arXiv:2605.04481v1 Announce Type: new Abstract: Minimum-fuel low-thrust rendezvous guidance yields bang-bang control structures highly sensitive to estimation errors, sensor anomalies, and solver regu

Transformation Categorization Based on Group Decomposition Theory Using Parameter Division

arXiv:2605.04056v1 Announce Type: new Abstract: Representation learning seeks meaningful sensory representations without supervision and can model aspects of human development. Although many neural ne

Tree-Conditioned Edit Flows for Ancestral Sequence Reconstruction

arXiv:2605.04119v1 Announce Type: cross Abstract: Ancestral sequence reconstruction (ASR) aims to infer extinct protein sequences at internal nodes of a phylogenetic tree. Classical ASR methods are ty

TSCG: Deterministic Tool-Schema Compilation for Agentic LLM Deployments

arXiv:2605.04107v1 Announce Type: cross Abstract: Production agent frameworks (OpenAI Function Calling, Anthropic Tool Use, MCP) transmit tool schemas as JSON, a format designed for machine parsing, n

UAV as Urban Construction Change Monitor: A New Benchmark and Change Captioning Model

arXiv:2605.04409v1 Announce Type: new Abstract: Remote Sensing Image Change Captioning (RSICC) aims to generate spatially grounded natural language descriptions of scene evolution from bi-temporal ima

UFAL-CUNI at SemEval-2026 Task 11: An Efficient Modular Neuro-symbolic Method for Syllogistic Reasoning

arXiv:2605.04941v1 Announce Type: new Abstract: This paper describes our system submitted to SemEval-2026 Task 11: Disentangling Content and Formal Reasoning in Large Language Models. We present an ef

Unified Framework of Distributional Regret in Multi-Armed Bandits and Reinforcement Learning

arXiv:2605.05102v1 Announce Type: new Abstract: We study the distribution of regret in stochastic multi-armed bandits and episodic reinforcement learning through a unified framework. We formalize a di

U.S. intelligence says Iran can outlast Trump’s Hormuz blockade for months — via @washingtonpost https://www.washingtonpost.com/national-sec…

model-releasesanthropic--x

U.S. intelligence says Iran can outlast Trump’s Hormuz blockade for months — via @washingtonpost https://www.washingtonpost.com/national-security/2026/05/07/cia-intelligence-iran-trump-blockade-missil

VCBench: Benchmarking LLMs in Venture Capital

arXiv:2509.14448v2 Announce Type: replace Abstract: Benchmarks such as SWE-bench and ARC-AGI demonstrate how shared datasets accelerate progress toward artificial general intelligence (AGI). We introd

Vibe launches Dot, a wearable AI device for capturing real-world workplace conversations

model-releasessiliconangle

Vibe Inc., the creator of a contextual artificial intelligence workspace platform that handles real-world meetings, introduced a wearable device today that brings an AI assistant to professionals wher

Vision-EKIPL: External Knowledge-Infused Policy Learning for Visual Reasoning

arXiv:2506.06856v3 Announce Type: replace Abstract: Visual reasoning is crucial for understanding complex multimodal data and advancing Artificial General Intelligence. Existing methods enhance the re

VL-UniTrack: A Unified Framework with Visual-Language Prompts for UAV-Ground Visual Tracking

arXiv:2605.04574v1 Announce Type: new Abstract: UAV-ground visual tracking (UGVT) aims to simultaneously track the same object from both the UAV and the ground view. However, existing two-stream metho

VTAgent: Agentic Keyframe Anchoring for Evidence-Aware Video TextVQA

arXiv:2605.04870v1 Announce Type: new Abstract: Video text-based visual question answering (Video TextVQA) aims to answer questions by reasoning over visual textual content appearing in videos. Despit

Wasserstein-Aligned Localisation for VLM-Based Distributional OOD Detection in Medical Imaging

arXiv:2605.05161v1 Announce Type: new Abstract: Zero-shot anomaly localisation via vision-language models (VLMs) offers a compelling approach for rare pathology detection, yet its performance is funda

We already had gemini-3.1-flash-lite-preview back on March 3rd, not clear if this new gemini-3.1-flash-lite is different other than no longe…

model-releasessimon-willison--x

We already had gemini-3.1-flash-lite-preview back on March 3rd, not clear if this new gemini-3.1-flash-lite is different other than no longer being marked as a 'preview'. Pricing appears to be the sam

We know you’re eager for voice updates in ChatGPT. Stay tuned, we’re cooking.

OpenAI announced on X that voice feature updates for ChatGPT are in development and coming soon, though no specific timeline was provided. The post acknowledges user demand for enhanced voice capabili

model-releasesopenai--x

We recently found some instances of CoT grading during the training of previously deployed models after building a system that scans all Ope…

We recently found some instances of CoT grading during the training of previously deployed models after building a system that scans all OpenAI RL runs for accidental CoT grading. We did not find clea

model-releasesopenai--x

Welcome to DS4, a specialized inference engine for DeepSeek v4 Flash. https://github.com/antirez/ds4 This project would have been impossible…

model-releasesjeremy-howard--x

Welcome to DS4, a specialized inference engine for DeepSeek v4 Flash. https://github.com/antirez/ds4 This project would have been impossible without the existence of llama.cpp and GGML and the work of

We've teamed up with @cerebras to offer free Windsurf plans for SWE-1.6 Fast Mode at up to 1000 tok/s! Fast Mode is built on Cerebras infere…

model-releaseswindsurf--x

We've teamed up with @cerebras to offer free Windsurf plans for SWE-1.6 Fast Mode at up to 1000 tok/s! Fast Mode is built on Cerebras inference, enabling superior speed for planning and development wi

What an amazing event! Thank you @AnthropicAI. Doctors’ coding community in full swing.

model-releasesboris-cherny--x

Boris Cherny praised an Anthropic AI event that brought together a doctors' coding community, expressing enthusiasm about the gathering's energy and engagement. The post suggests Anthropic hosted or s

What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis

arXiv:2605.03354v1 Announce Type: new Abstract: Agent memory failures are silent: an LLM-based agent can produce a fluent response even when it fails to extract, retain, or retrieve the information ne

While Anthropic will use the Colossus 1 data center, which has a really bad environmental record, xAI retains the larger Colossus 2 for its own AI training (Simon Willison/Simon Willison's Weblog)

Simon Willison / Simon Willison's Weblog: While Anthropic will use the Colossus 1 data center, which has a really bad environmental record, xAI retains the larger Colossus 2 for its own AI training —

model-releasestechmeme

who’s adding this to reachy mini?

model-releasesclem-delangue--x

who’s adding this to reachy mini? Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents. Voice agents are now real-time collaborat

With the help of Claude Mythos Preview, the Firefox team fixed more security bugs in April than in the past 15 months combined.

model-releasesboris-cherny--x

The Firefox development team reportedly resolved a significantly higher volume of security vulnerabilities in April with assistance from Claude Mythos Preview, an AI tool, compared to their bug-fixing

Wordle 1,782 3/6 ⬛⬛⬛⬛🟨 🟩🟩🟨⬛⬛ 🟩🟩🟩🟩🟩

model-releasesanthropic--x

This post shows a Wordle game result where the player solved puzzle #1,782 in 3 attempts, displaying the emoji grid that represents their guesses and letter placements (gray for incorrect letters, yel

Wordle 1,783 6/6 ⬛⬛⬛⬛🟩 ⬛⬛⬛🟩⬛ ⬛⬛🟩🟩🟩 ⬛🟩🟩🟩🟩 ⬛🟩🟩🟩🟩 🟩🟩🟩🟩🟩

model-releasesanthropic--x

This post documents a Wordle game completion (puzzle #1,783) where the player solved the word in six attempts, the maximum allowed before failure. The color-coded emoji grid shows the progression of g

12 Angry AI Agents: Evaluating Multi-Agent LLM Decision-Making Through Cinematic Jury Deliberation

arXiv:2605.01986v1 Announce Type: new Abstract: What if the twelve jurors of Sidney Lumet's 12 Angry Men (1957) were not men, but large language models? Would the one juror who disagrees still be able

6G Needs Agents: Toward Agentic AI-Native Networks for Autonomous Intelligence

arXiv:2605.01546v1 Announce Type: cross Abstract: Sixth-generation (6G) networks are increasingly envisioned as AI-native infrastructures integrating communication, sensing, and computing into a unifi

A Benchmark for Interactive World Models with a Unified Action Generation Framework

arXiv:2605.03941v1 Announce Type: new Abstract: Achieving Artificial General Intelligence (AGI) requires agents that learn and interact adaptively, with interactive world models providing scalable env

A critical question in agent design is “how do we build agentic workflows so humans are given significant, interesting, or variance-producin…

model-releasesethan-mollick--x

A critical question in agent design is “how do we build agentic workflows so humans are given significant, interesting, or variance-producing decisions as they come up in the work?” A Claude-run compa

A Deeper Dive into the Irreversibility of PolyProtect: Making Protected Face Templates Harder to Invert

arXiv:2605.03857v1 Announce Type: new Abstract: This work presents a deeper analysis of the 'irreversibility' property of PolyProtect, a biometric template protection method initially proposed for sec

A Domain Incremental Continual Learning Benchmark for ICU Time Series Model Transportability

arXiv:2605.03832v1 Announce Type: new Abstract: In recent years, machine learning has made significant progress in clinical outcome prediction, demonstrating increasingly accurate results. However, th

A Framework for Exploring and Disentangling Intersectional Bias: A Case Study in Fetal Ultrasound

arXiv:2605.02942v1 Announce Type: cross Abstract: Bias in medical AI is often framed as a problem of representation. However, in image-based tasks such as fetal ultrasound, performance disparities can

ABD: Default Exception Abduction in Finite First Order Worlds

arXiv:2602.18843v3 Announce Type: replace Abstract: We introduce ABD, a benchmark for default-exception abduction over finite first-order worlds. Given a background theory with an abnormality predicat

AcademiClaw: When Students Set Challenges for AI Agents

arXiv:2605.02661v1 Announce Type: new Abstract: Benchmarks within the OpenClaw ecosystem have thus far evaluated exclusively assistant-level tasks, leaving the academic-level capabilities of OpenClaw

Adaptive Negative Scheduling for Graph Contrastive Learning

arXiv:2605.03076v1 Announce Type: new Abstract: Graph contrastive learning (GCL) has become a central paradigm for self-supervised representation learning in computational intelligence, with applicati

AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages

arXiv:2601.06395v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly multilingual, yet open models continue to underperform relative to proprietary systems, with the gap m