Instead of treating 3D reconstruction and video generation as standalone alternatives, ArtiFixer combines their strengths. > Phase I: Trains a powerful bidirectional generative video model to transport degraded renderings into clean frames.
> Phase II: Distills the teacher model
RESEARCH
-
ArtiFixer combines 3D reconstruction and video generation in two phases
By
–
-
NVIDIA Research’s ArtiFixer uses video diffusion to fix 3D reconstruction artifacts
By
–
NVIDIA Research just dropped a game-changer for 3D reconstruction.
— AlphaSignal (@AlphaSignalAI) 23 juin 2026
If you've ever dealt with blurry holes or floating artifacts in 3D Gaussian Splatting, ArtiFixer is the solution.
It uses a powerful auto-regressive video diffusion model to seamlessly repair and extend 3D… pic.twitter.com/5GLByeqf5PNVIDIA Research just dropped a game-changer for 3D reconstruction. If you've ever dealt with blurry holes or floating artifacts in 3D Gaussian Splatting, ArtiFixer is the solution. It uses a powerful auto-regressive video diffusion model to seamlessly repair and extend 3D
-
PlanBench-XL: Evaluation of Long-Horizon Planning of LLM Agents
By
–
PlanBench-XL Evaluation of Long-Horizon Planning of Tool-Using LLM Agents in Large-Scale Tool Ecosystems
-
Auto-Harness: Self-Improving Harnesses
By
–
New paper on agents that shape their own harnesses to improve over time. No LangChain, but built on DeepAgents! Three key steps: 1/ Exploitation of weaknesses: identify the modes of
-
OpenRouter to solve GPU shortage in research
By
–
Tired of seeing “out of capacity” while looking for GPUs? Introducing OpenRouter for compute
— alphaXiv (@askalphaxiv) 23 juin 2026
In the past 6 months, overwhelming demand for compute has made pricing and availability huge challenges for the research community
We aggregate compute offerings from across the… pic.twitter.com/Z3IdyPvbkmTired of seeing "capacity exhausted" when searching for GPUs? Introducing OpenRouter for Compute Over the past 6 months, overwhelming demand for compute has made pricing and availability real challenges for the research community We aggregate
-
Krea 2 Raw: Undistilled Model for Further Development
By
–
Krea 2 Raw is an undistilled model from mid-training stage that we created while training Krea 2 Medium. The lack of distillation, fine-tuning, or post-training make it an ideal model to build on.
-
Krea 2 open weights: Raw and Turbo models released
By
–
today, we release the open weights of Krea 2.
— Krea (@krea_ai) 23 juin 2026
welcome Krea 2 Raw and Krea 2 Turbo, an undistilled model from mid-training meant to be fine-tuned, and a fast distilled version with a wide aesthetic diversity.
read the details below 👇 pic.twitter.com/3ymzUL2bxvtoday, we release the open weights of Krea 2. welcome Krea 2 Raw and Krea 2 Turbo, an undistilled model from mid-training meant to be fine-tuned, and a fast distilled version with a wide aesthetic diversity. read the details below
-
Testing AI bug detection by editing out bugs for false positive check
By
–
Testing this more. I think it's doing the right thing based on the image rather than prior knowledge. In these two tests I edited the original image to remove the bugs while keeping everything else the same. I was looking for a false positive if it wasn't doing the correct
-
AI Search Default: Manus Automates SEO Workflows, Finds 8,456 Slow Pages
By
–
AI search is the default now, but SEO didn't die, it leveled up. Manus rolls keyword research, audits, competitor analysis & backlink prospecting into one workflow. Describe the goal, it runs it. One sample audit flagged 8,456 slow pages on a 180k/day site.
-
Reinforcement learning for broadly persistent beneficial models
By
–
Reinforcement learning towards broadly and persistently beneficial models
by Akshay V. Jagadeesh Rahul K. Arora @OpenAI Learn more: https://
bit.ly/4eAqbsm #AI #GenerativeAI #ArtificialIntelligence #MachineLearning