https://zlio.com/AI news, deduplicated.https://huggingface.co/blog/allenai/hiro-acev2 zlio-3902Wed, 14 Oct 2026 16:12:42 GMT HiRO-ACEv2 is an AI weather model that simulates extreme weather at 3 km resolution for a fraction of the usual cost, released on Hugging Face.https://www.reddit.com/r/LocalLLaMA/comments/1x5tsni/inclusionailing31flash_hugging_face zlio-3925Wed, 14 Oct 2026 15:11:14 GMT An open Hugging Face model with 560B total parameters, 25B active per token and a 1M-token context window. It follows Ling 3.0, which was 127B.https://huggingface.co/papers/2610.17953 zlio-3359Tue, 13 Oct 2026 00:00:00 GMT Agent-Reflex is a training framework that teaches language agents to reflect on noisy MCP tool feedback and user corrections, improving failure recovery in long-horizon tasks.https://huggingface.co/papers/2610.18774 zlio-3466Tue, 13 Oct 2026 00:00:00 GMT GreenWAM is a world foundation model that both predicts how a robot's actions change a scene and chooses those actions, aiming to work across different robot bodies.https://huggingface.co/papers/2610.18868 zlio-3810Tue, 13 Oct 2026 00:00:00 GMT Paper Tev presents a method to scale long-horizon agentic training environments using dense progress rewards, aiming to help open models work autonomously for hours to days.https://www.reddit.com/r/StableDiffusion/comments/1x52tz3/testing_texturefixvaeforqwenimage21 zlio-2877Tue, 13 Oct 2026 17:23:44 GMT A VAE that reduces checkerboard artifacts in Qwen Image 2.1 generations, shared on Hugging Face. Differences are only visible when zooming past 500%.https://www.reddit.com/r/StableDiffusion/comments/1x52j87/aziib_pixel_style_lora_pixel_art_lora_for_krea2 zlio-3389Tue, 13 Oct 2026 17:13:00 GMT A pixel-art LoRA for image models Krea2 and Anima, downloadable from Civitai with the trigger word 'aziib pixel style'.https://huggingface.co/papers/2610.18356 zlio-2466Tue, 13 Oct 2026 00:00:00 GMT Urbex is a research framework that lets a vision-language agent actively search a city-scale 3D scene, querying landmarks and zooming instead of ranking precomputed instance proposals.https://huggingface.co/papers/2610.18615 zlio-2468Tue, 13 Oct 2026 00:00:00 GMT VideoJev is a family of models that turn video understanding and interaction into structured decision tasks, adapting pretrained MLLMs with a shared prediction head instead of autoregressive text answers.https://huggingface.co/papers/2610.18773 zlio-2469Tue, 13 Oct 2026 00:00:00 GMT Jeva is a paper arguing current system-one model designs miss four structural requirements, including distinguishing answer absence from uncertainty and permutation invariance, and proposes models built around them.https://huggingface.co/papers/2610.18568 zlio-3649Tue, 13 Oct 2026 00:00:00 GMT A research paper proposing lightweight latent future prediction to connect vision-language-action policies with world action models, aiming to model action consequences without costly video generation.https://huggingface.co/papers/2610.17928 zlio-2467Tue, 13 Oct 2026 00:00:00 GMT A paper examines how GUI agents should choose the modality for retrieved memory, arguing existing agents use a fixed representation instead of adapting the format to the task.https://huggingface.co/papers/2610.18871 zlio-3648Tue, 13 Oct 2026 00:00:00 GMT A research paper on sparse-to-dense motion completion: recovering dense mesh motion from sparse vertex displacements. It shows residual priors trained on one geometry family often fail to beat parameter-free harmonic extension on other families.