The top 20 hugging face models & datasets measured by RiseFinder over the 30d window ending November 2026.
EgoStandard is a video dataset from LightwheelAI. It is described as the 90,000-hour head-view line of EgoSuite-Open100K. The dataset pairs head-view egocentric video with synchronized 3D hand pose, and its body subset adds full-body pose. It is distributed through the LightwheelAI/EgoStandard Bucket, with a Git repository serving as the dataset card and access point for downloads.
Hugging Face dataset
Open source ↗This is an experimental character-replacement LoRA. It was trained for 1,000 updates and aims to replace a selected character using an image reference. The process is designed to preserve the source scene during the replacement. In local comparisons, it often preserved the background more closely than the base model.
Hugging Face model
Open source ↗venvoo/china-a-share-l2-level2-limit-order-book-tick-data is a Hugging Face dataset containing a historical archive of Chinese exchange Level-2 market data from 2017 to 2026, supplied through a vendor export and stored in Parquet format. It includes ten-level quote snapshots, individual order messages, and trade-stream records, covering A-share stocks as well as non-stock instruments such as ETFs and bonds.
Hugging Face dataset
Open source ↗This is a conversational speech dataset. Basis Conversations 1500 contains multi-party, multilingual, full duplex conversations with up to four simultaneous speakers per conversation, each recorded on channel-separated 48 kHz audio. Conversations seat between two and four people, the median conversation lasts 33 minutes, and 2,645 unique speakers are represented.
Hugging Face dataset
Open source ↗This is a multimodal decision classifier designed to process various data types including text, images, audio, and video. Users supply a question along with specific options. The model then outputs a probability for each provided option, providing a decision rather than generating a detailed explanation.
Hugging Face model
Open source ↗AxiomicLabs/Tiny_Theory_of_Mind is a benchmark dataset for evaluating theory of mind capabilities in small language models. It is described as its creators' first attempt at this kind of evaluation. The benchmark covers a wide variety of theory of mind topics at difficulty levels that, for humans, would be appropriate for Pre-K through 6th grade.
Hugging Face dataset
Open source ↗Mitsuba-ComfyUI-27B-GGUF is a text generation model distributed in GGUF format on Hugging Face. It is a ternary (1.58-bit) quantized version of Qwen3.8-27B, tuned for use with ComfyUI. Its stated purpose is writing image and video generation prompts that follow strict conditions, and describing images.
Hugging Face model
Open source ↗Vyber07/cyber-security is a cybersecurity knowledge base dataset intended for AI training. It covers cybersecurity domains including offensive red teaming, bug bounty exploitation, defensive SOC operations, digital forensics, and AI/LLM security. The dataset is 16 GB in size, contains 507 files, and spans more than 30 domains drawn from more than 15 platforms.
Hugging Face dataset
Open source ↗GLiNER2.5-Decide is a 340M-parameter English text classification model in the GLiNER2.5 family. Instead of using a fixed prompt template, it accepts any label set at call time, covering uses such as intent, routing, sentiment, priority, policy, and multi-label tagging, all in a single forward pass without generating tokens. It can be loaded with AutoExtractor and run locally.
Hugging Face model
Open source ↗Julia 1 is a machine learning model from SupersonicLabs and the first model in its Julia family. It is a small model of 144.3M parameters that takes a state, a question, and possible answers as input and produces a single decision. The same interface supports classification, routing, ordered scores, and Boolean decisions.
Hugging Face model
Open source ↗IQuest-Q1 is a Mixture-of-Experts (MoE) language model developed by IQuest for agentic coding, reasoning, and multi-step tool use. It has approximately 320B total parameters with about 15B activated per token, exposes an OpenAI-compatible API, and can be deployed in production with SGLang or vLLM. The model accepts text-only input and its documentation notes that generated code and explanations can be incorrect and may require human oversight.
Hugging Face model
Open source ↗Open-Jev is a dataset that pairs a state and a question with a typed decision output. Depending on the task, that decision can take the form of a yes/no probability, a distribution over choices, independent label probabilities, or a discrete numeric or ordinal decision. The repository publishes twelve separate, frozen data configs from the Open-Jev project.
Hugging Face dataset
Open source ↗fastino/fast-decisions is a classification dataset suite associated with GLiNER2.5-Decide. It covers seventeen domains, with one file per domain. Each row contains a document plus the decisions a product has to make, including the task name, the candidate labels, and the gold label. The suite includes both single-label heads, which have one gold string, and multi-label heads.
Hugging Face dataset
Open source ↗bones-studio/seed, also called BONES-SEED (Skeletal Everyday Embodiment Dataset), is an open dataset of 142,220 annotated human motion animations intended for humanoid robotics. It contains motion capture data in SOMA and Unitree G1 formats, along with natural language descriptions, temporal segmentation, and detailed skeletal metadata.
Hugging Face dataset
Open source ↗This is a dataset release on Hugging Face. It publishes 326,619 typed decision records, including 147,139 training records, drawn from the frozen community-hard-mix-v2-final mixture used for the Open-Jev 27B training stage. It is a redistributable projection rather than the exact complete training or evaluation dataset.
Hugging Face dataset
Open source ↗TIGER-Lab/MMLU-Pro is a benchmark dataset for evaluating large language models. It is a massive multi-task understanding dataset designed to test model capabilities more rigorously. It contains 12K complex questions spanning various disciplines, which models answer so their performance can be measured.
Hugging Face dataset
Open source ↗Cirquar Tech/wm_imagined is a Hugging Face dataset of imagined robot interaction segments generated by a world model. Its current release contains only the RoboTwin2.0 subdataset, covering 50 robot tasks. Each HDF5 chunk pairs 21 observation frames from three camera views with 20 action steps, rewards, and termination flags. The publisher documents additional subdatasets as planned, not already included in this release.
Hugging Face dataset
Open source ↗MiMo-V2.6-Distill-Qwen-9B is a 9-billion-parameter agentic language model developed by Xiaomi MiMo. It was created through supervised fine-tuning of Qwen3.5-9B on MiMo-generated data and covers coding, general-purpose agent tasks, visual coding, and cybersecurity. It is released as an SFT checkpoint intended as a starting point for open research in agent work.
Hugging Face model
Open source ↗markov-ai/cad-1000-hours is a dataset of 509 end-to-end, single-display Windows CAD task recordings across seven CAD software families, totaling 1,018.1229 hours. Each task includes a task prompt, reference input files, submitted CAD deliverables, task-specific evaluation rubrics, a screen recording video, frame-synchronized events, metadata, and frame-level task narration. The dataset contains 5,455 public task objects totaling 207.18 GiB, and every task includes all seven required root files plus at least one input and one output.
Hugging Face dataset
Open source ↗Anthropic/claude-protein-binder-design is a dataset of protein design experiment results. It contains 1,440 de novo miniprotein binders of 50 to 120 residues designed against 16 targets by two Claude models operating as autonomous protein-design agents, with wet-lab characterization from two contract research organizations. Each design includes binding calls and kinetics, structure predictions, and design provenance, and the tables join on a uuid column. Data and documentation are licensed CC BY 4.0.
Hugging Face dataset
Open source ↗Get early signals on the websites, apps, repositories and startups gaining momentum. Join the list for Risefinder email briefings.
By subscribing, you ask us to store your email and send Risefinder updates. You can unsubscribe at any time. Privacy.
Want a sample first? Read past daily emails.