AN IN-DEPTH REPOSITORY
Every model, paper, media release, agent and tool from the month—in one source-linked archive.
COMPLETE MONTHLY ARCHIVE
Control became the creative advantage: better video direction, longer-running agents and open systems for science and robotics.
THE COMPLETE MONTHLY INDEX
Every linked paper, product release, research project and article has been checked directly. Cards use the publication's own description, imagery and published date wherever the source supplied them.
The AI Search helps discover the wider field, while every available card opens the original publication. Dates come only from publication metadata or official source APIs—never from a video. A checked date records our review, not the release. Older undated entries show their archive month.
MINIMAX.IOMiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model - MiniMax Research
M3 reaches frontier capability on coding and agentic tasks, introduces the brand-new MSA (MiniMax Sparse Attention) supporting up to 1M context, and is a natively multimodal model. It is the only domestic model combining all three Frontier essentials and…
AUTOSCIENTISTS.OPENSCIENTIST.AIAutoScientists — Self-Organizing Agent Teams for Long-Running Scientific Experimentation
A decentralized team of AI agents that self-organize into research teams, run experiments in parallel, and reorganize when directions stagnate. +8.33% on BioML-Bench, 1.9× faster on GPT optimization, +12.5% Spearman on ACE2–Spike.
YFYANG007.GITHUB.IOControlLight: Towards Controllable, Consistent, and Generalizable Low-Light Enhancement
DEEPSWE.DATACURVE.AIDeepSWE measures frontier coding agents on original, long-horizon software engineering tasks.
RESEARCH.NVIDIA.COMγ-World: Generative Multi-Agent World Modeling Beyond Two Players
γ-World: A generative multi-agent world model for interactive simulation
KASOTHAPHIE.GITHUB.IOGenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction
High-fidelity PBR mesh reconstructions of 3D indoor scenes from sparse images or casual videos leveraging a 3d generative prior.
InstructAV2AV: Instruction-Guided Audio-Video Joint Editing
Project page for InstructAV2AV: Instruction-Guided Audio-Video Joint Editing.
RESEARCH.NVIDIA.COMParallel Box Decoding unlocks substantial parallelism and preserves intra-box geometric coherence.
HUGGINGFACE.COopenbmb/MiniCPM5-1B · Hugging Face
MiniCPM5 1B is published on Hugging Face. Open the model or project card for the author’s documentation, files, license and usage details.
ANTHROPIC.COMIntroducing Claude Opus 4.8
Our latest model, Claude Opus 4.8, is an upgrade to our Opus class of models, with stronger performance across coding, agentic tasks, and professional work, and the consistency to handle long-running work.
KOI953215.GITHUB.IOPantheon360: Taming Digital Twin Generation via 3D-Aware 360° Video Diffusion
A 3D-aware 360° video diffusion model that generates controllable, geometrically consistent panoramic videos from sparse inputs along user-defined camera trajectories. CVPR 2026.
PHYSX-OMNI.GITHUB.IOPhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects
PhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects
RESEARCH.NVIDIA.COMPiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion
PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion
MLFARINHA.GITHUB.IOPIXLRelight: Controllable Relighting via Intrinsic Conditioning
PIXLRelight: Controllable Relighting via Intrinsic Conditioning.
VCAI.MPI-INF.MPG.DERelightable Holoported Characters: Capturing and Relighting Dynamic Human Performance from Sparse Views
We present Relightable Holoported Characters (RHC), a novel person-specific method for free-view rendering and relighting of full-body and highly dynamic humans solely observed from sparse-view RGB videos at inference. In contrast to classical…
SCOPE: Simulating Cross-game Operations in Playable Environments
SCOPE: Simulating Cross-game Operations in Playable Environments for FPS World Models
SEGA
SEGA: Spectral-Energy Guided Attention for Resolution Extrapolation in Diffusion Transformers
GUOWEIXU.COMSelf-Improving Language Models with Bidirectional Evolutionary Search
A search framework that couples forward candidate evolution with backward goal decomposition.
STATIC.STEPFUN.COMStep 3.7 Flash — A high-efficiency Flash model for Real-World
Multimodal Understanding & Action｜Web & Visual Search Enhancement｜Reliable Tool Use & Orchestration｜Agent Ecosystem Compatibility
LHMD.TOPTriSplat | Simulation-Ready Feed-Forward 3D Scene Reconstruction
TriSplat: Simulation-Ready Feed-Forward 3D Scene Reconstruction project page.
ELEVENLABS.IOIntroducing Dubbing v2
ElevenLabs’ new dubbing model conditions on the original performance to carry tone, pacing and emotional intent into synchronized speech across more than 90 languages.
RAI-INST.COMA New Benchmark in Robot Juggling | RAI Institute
AthenaZero can juggle barehanded using onboard vision feedback. By using multi-fingered hands, it can transition seamlessly between a wide range of juggling patterns.
PRISMML.COMPrismML — Introducing 1-bit and Ternary Bonsai Image 4B: Image Generation for Local Devices
Today we’re releasing Bonsai Image 4B , a family of compact image-generation models designed to run high-quality diffusion inference on local hardware: from laptops to phones.
ELEVENLABS.IOIntroducing Music v2, our groundbreaking new music model
Music v2 improved vocals, instrumentation, multilingual output and arrangement while adding section-level composition and inpainting for regenerating only part of a track.