The paper explores how LLM agents use test-time compute to improve solutions, introducing Elo-per-token analysis to measure performance scaling. It finds that agents initially outperform independent sampling but eventually lag behind, while humans show sustained improvement.
The chart will appear after repeat observations. The current metric comes from the source.
5 source votesReal observations only. History before source connection is not reconstructed.
This research provides insights into how LLM agents utilize computational resources during testing, helping to optimize their performance and understand their limitations.
The paper introduces Vidu S2, a system with two components: Vidu S2-Avatar for real-time interactive digital characters and Vidu S2-Editing for real-time video editing. It also explores real-time spatial video generation for both components. Compared to its predecessor, Vidu S2-Avatar offers improved real-time 720p video generation, dynamic reference updates, and enhanced instruction following. Vidu S2-Editing enables real-time video stream editing with features like style rendering and background replacement.
5 days agoDaily PapersPhysBrain 1.5 is a unified model for understanding physical environments, generating actions, and predicting future states. It uses a vision-language foundation, encodes language responses and motion as sequences, and is pre-trained on human interaction videos. The model achieves strong performance on embodied understanding benchmarks.
yesterdayDaily PapersZGCM-1 is a 7B parameter foundation model designed for math and agentic search, emphasizing efficiency through architectural and system co-design, progressive curriculum learning, and an AI-native R&D workflow. It demonstrates competitive performance against larger models on mathematical and search tasks, with improvements in training efficiency and open-source availability of training materials.
4 days agoDaily PapersAtria Dawn Preview is a foundation agentic language model designed for scientific research and engineering workflows, trained via a Verifiable Experience Pipeline that connects tool-mediated interactions to executable environments and externally verified outcomes. It achieves competitive performance across 16 benchmarks and shows potential for human-AI collaboration in research and development.
yesterdayDaily PapersReactHuman is a benchmark for evaluating how well embodied multimodal large language models (MLLMs) can make quick, human-like decisions in response to sudden physical dangers. It includes over 1,000 reproducible scenes with precise physics-based simulations and measures reactions based on reasonableness, safety, and physical accuracy.
6 days agoDaily PapersAlayaVista is a streaming video world model that separates panoramic world evolution from perspective observation synthesis. It uses a pretrained panorama expansion model to create a 360-degree scene prior and evolves it as a camera-conditioned panoramic latent state, then maps this to the requested perspective video latents.
2 days ago