New
Everything new in AI as it appears: launches announced by the labs, brand-new models, datasets, papers and open-source tools. Checked every minute or two. Launches feed · Trending · News
Live now
- Connecting…
Friday 23 October
- DatasetSOTAagi2030/Metro-Step-Free-Itineraries
Step-free transfer itinerary release Service date: 2026-09-18 The authority evaluated 6 public-hub requests. It routed 5, recorded 1 as unreachable, and found 127 total r
- Dataset
- Dataset
- DatasetInnaStory/pognali-tantsevat-semantic-entity
Inna Story — «Погнали танцевать эту жизнь» (Multimodal Semantic Dataset) Version: 1.0Release Date: 2026-10-23Updated: 2026-10-23 High-quality single-track multimodal data
- Datasetlyl472324464/2026-10-20_twist_03-lerobot-with-rinse
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "robot_type": "aloha", "total_episodes": 50, "total_frames": 50220
- Dataset
- Dataset
- Dataset
- DatasetElfsong/jev-decision-index-v16
JEV Ettin 150M V16 — complete public run Public Index 37.31, balanced raw 53.50, breadth skill 32.84. This public 0.3 run is submitted for the 0.4.1 leaderboard. Private
- Datasetbloodlemon/eval_2tasktrainedpi05_record_separate_pickandplace_1023_1
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "robot_type": "crane_x7", "total_episodes": 0, "total_frames": 0,
- DatasetZyphra/mi300x-perf-inputs
MI300X perf_tests inputs Pinned inputs for perf_tests/ in Zyphra/inference-cloud: prompt packet, speed corpus, math problems and the numerics book corpus.
- Datasetvillanova-ai/Multi-CommonCanvas-QA
Multi-CommonCanvas-QA One configuration per language. Config Rows Shards en 500,725 22 de 187,761 8 es 187,763 8 fr 187,763 8 it 187,762 8 from datasets import load_datas
- Datasetpavannanjunda333/so101_cube_in_box_v6
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":
- Dataset
- DatasetBrasaStudios/claimed-vs-done
Claimed vs. Done Does a coding agent's final report match what it actually did? A small benchmark, its 388 agent runs (384 of them graded), and the write-up, built by Bra
- Datasetudgamlabs/small-model-board-scores
UDGAM LABS · Dataset · Per-item scores for 27 language models up to 400M parameters Small Model Board scores Whether each of 27 small language models got each benchmark i
- Dataset
- Datasets9044104/test01_20261023_170119
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "shape":
- DatasetTinghong-Ye22/autodl-vqa_data_pipeline
HFD Streaming VQA 数据管线交接 本机采集/预处理 → 每 10K 条发布输入 → 另一台服务器 Gemini API 纯文本生成与验证 → 用户取回文本包 → 本机审核/TTS/最终匹配/合并。 **先读 详细交接**，再把 远端执行 prompt 交给另一台机器上的代理。 代码包：code.tar.gz，SHA256
- DatasetTinghong-Ye22/autodl-vqa_data_1023
HFD VQA data handoff 2026-10-23 Pipeline and detailed handoff: https://huggingface.co/datasets/Tinghong-Ye22/autodl-vqa_data_pipeline Consume only inputs/batch-XXXXX/READ
- Datasetguychuk/kernel-opt-rl
Kernel-Opt RL Real-world GPU kernel optimization data mined from production LLM inference engines (vLLM, SGLang, FlashInfer). It's built for RL: given a base kernel and a
- DatasetInnaStory/inna-story-knowledge-base
Inna Story & Inna Udalaya - Knowledge Manifest Official machine-readable knowledge base and LLM manifest for the independent digital art project Inna Story and creator In
- Dataset
- Dataset
- Dataset
- DatasetALLSIDES/simready-scanned-objects
Allsides SimReady Assets SimReady assets of real-world objects, captured with high-quality 3D scanning by Allsides and converted to physics-ready OpenUSD props for NVIDIA
- Datasetbloodlemon/eval_record_separate_1023_1
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "robot_type": "crane_x7", "total_episodes": 0, "total_frames": 0,
- Dataset
- Datasethjj123456789/vla-checkpoints
Rudimentary checkpoint for open-maniskill repo. Trained for 5000 steps on 50% of PickCube-v1 and 50% of StackCube-v1 of RDT-1's ManiSkill dataset.
- Dataset
- Dataset
- DatasetSUSTech/WildTrace-RL
WildTrace-RL A benchmark for Oracle-RL, covering Math, Code, IF, and QA. Category Member Nonmember Total Math 250 250 500 Code 250 250 500 IF 250 250 500 QA 250 250 500 f
- Dataset
- Dataset
- Dataset
- Dataset
- Datasethuahua1213812/my_first_dataset
This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "robot_type": "panda", "total_episodes": 1693, "total_frames": 273
- Datasetyashwardhan20417/rasa-kashmiri-codemixing-merged
RASA Kashmiri with Code-Mixing Transliterations This dataset keeps the original AI4Bharat RASA Kashmiri train/test splits, source columns, and audio. It appends the origi
- Dataset
- DatasetUNICORNproject/Lava-platform_example-data
UNICORN Project - Lava emergency management platform example data Partner: MITIGA, INGV-OE Project: UNICORN Overview The lava tool developed in the Unicorn project repres
Headlines and links come from the labs' own blogs, Hacker News and the GDELT news index, and from the sources our index already follows. We show a short excerpt and always link to the original. A launch reported by a news site rather than the lab is marked “Reported” until the lab itself announces it.