Sign inGet started

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…

Yesterday

  • Dataset
    yagmurbozatli/kivicik-1.5b-veriyagmurbozatli · huggingface · 10h ago
  • Dataset
    starpilot-ai/mobile_umi_test_3

    This dataset was created using LeRobot. Dataset Description One take of egocentric bimanual manipulation, split into three episodes: episode task frames duration 0 calibr

    starpilot-ai · huggingface · 10h ago
  • Dataset
    Zeteng/meva-reversible-events-vjepa2.1-vitb-features

    MEVA reversible events — frozen V-JEPA 2.1 ViT-B features Pre-computed V-JEPA 2.1 ViT-B/16 (384) features for 10 pairs of reversible activities from the MEVA dataset, for

    Zeteng · huggingface · 10h ago
  • Dataset
    thekevinscott/surprisals-experiment-runsthekevinscott · huggingface · 10h ago
  • Dataset
    Viviannnyang/COIN-caption-action-temporal-8565

    COIN Caption, Action and Temporal Annotations This dataset combines four non-overlapping COIN video cohorts: 8,565 videos and 8,565 captions, built from 30,507 retained o

    Viviannnyang · huggingface · 10h ago
  • Dataset
    Viviannnyang/COIN-caption-action-temporal-4173

    COIN Caption, Action and Temporal Annotations This release contains 10,544 original COIN step phrases from 4,173 videos, with one extracted action per phrase. Original so

    Viviannnyang · huggingface · 10h ago
  • Dataset
    ag10101/cloth-folding-episode-0

    This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 50, "features": { "observation.state": { "dtype": "float32"

    ag10101 · huggingface · 10h ago
  • Dataset
    SOTAagi2030/Northline-Buoy-Quality-Release

    Northline buoy quality release This calibrated extract covers 2 active buoys. It contains 7 selected metrics, 1 missing required metrics, and 3 threshold alerts. Readings

    SOTAagi2030 · huggingface · 10h ago
  • Dataset
    rimur01/so101-lab5_20261022_040624

    This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":

    rimur01 · huggingface · 10h ago
  • Dataset
    xsfera/ornith-benchxsfera · huggingface · 10h ago
  • Dataset
    mengct00/sm_vec.de-en.cosyvoice3.1.4mengct00 · huggingface · 10h ago
  • Dataset
    krumphau/dbr-german-catholic-counterskrumphau · huggingface · 10h ago
  • Dataset
    artemZholus/so101_pick_cube

    This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":

    artemZholus · huggingface · 10h ago
  • Dataset
    plantwaterer/muria-decision-index

    muria: Decision Index 0.3 public-suite run The run files of plantwaterer/muria on the full public suite of the Decision Index, edition 0.3 (the public suite of the 0.4 bo

    plantwaterer · huggingface · 10h ago
  • Dataset
    krumphau/dbr-swedish-early-30yw-counterskrumphau · huggingface · 10h ago
  • Dataset
    alex2awesome/lipinski-tacit-knowledge

    Lipinski Rule of 5: papers and design-rationale extractions (private) Medicinal-chemistry and drug-discovery papers that cite or mention Lipinski's Rule of 5, or whose co

    alex2awesome · huggingface · 10h ago
  • Dataset
    diyheadshots/headshotdb

    HeadshotDB Open reference data on professional headshots from HeadshotDB (DIY Headshots): what 6,615 professionals across 6 professions wear and use in their directory he

    diyheadshots · huggingface · 10h ago
  • Dataset
    zehen8716/nyxozehen8716 · huggingface · 10h ago
  • Dataset
    ctxnn1/reipaint-paintings

    reipaint paintings Every painting from the reipaint experiments: Qwen3-8B writing p5.brush JavaScript that paints watercolours. Each row is a rendered 600 × 600 PNG with

    ctxnn1 · huggingface · 10h ago
  • Dataset
    hi-todayis-jh/ouro-1.4b-thinking-math-verl-per-context-rb-ragged-c1024-unb-ds4k-154711-20261020-rollouts

    Training rollouts: verl-per_context_rb-ragged-c1024-unb-ds4k-154711-20261020 The sampled responses of the verl_trainer run verl-per_context_rb-ragged-c1024-unb-ds4k-15471

    hi-todayis-jh · huggingface · 10h ago
  • Dataset
    Ollieboy/EquiDx-Clinical-Eval-Qwen

    Resource-Blindness in Frontier Clinical Triage Models 1. Identifying the Blind Spot: The "High-Resource" Assumption in AI Diagnostics Drawing from my background in Bioche

    Ollieboy · huggingface · 10h ago
  • Dataset
    aicps/cup_on_hotpad_20261021_113407

    This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":

    aicps · huggingface · 10h ago
  • Dataset
    SOTAagi2030/Borealis-Scan-Rights-468

    Borealis scan-rights authority Controlled permissions records for the Borealis derivative release desk.

    SOTAagi2030 · huggingface · 10h ago
  • Dataset
    SOTAagi2030/Borealis-Derivative-Queue-468

    Borealis derivative queue Controlled publication requests for an ice-core scan release exercise.

    SOTAagi2030 · huggingface · 10h ago
  • Dataset
    Saagara-Manju-Baiju/smb_act_pick_and_place_ds_20261021_234446

    This dataset was created using LeRobot. Dataset Structure meta/info.json: { "codebase_version": "v3.0", "fps": 30, "features": { "action": { "dtype": "float32", "names":

    Saagara-Manju-Baiju · huggingface · 10h ago
  • Dataset
    tc-huang/mahjong_v1_standrack01_redtc-huang · huggingface · 10h ago
  • Dataset
    cjc0013/celestial-holography-ai-research-candidate-v1

    Celestial Holography Under Exact Support Rigidity, soft distributions, amplitude flows, and nonlinear Lorentzian completion from a unified mathematical state This release

    cjc0013 · huggingface · 10h ago
  • Dataset
    skonan/PaperScorer-Vision-ICLR-27-5k-sample

    PaperScorer-Vision-ICLR-27-5k-sample A random sample of 5,000 ICLR 2027 submissions, in the same row format as skonan/PaperScorer-Vision (openreview-iclr config), scored

    skonan · huggingface · 10h ago
  • Dataset
    RandomThingsIDo/tinyclean-detmold-rirsRandomThingsIDo · huggingface · 10h ago
  • Dataset
    azazelovesH100/synthetic-aviation-lease-drive

    Synthetic Aviation Lease Drive (India, fictional) SYNTHETIC: fictional parties, not real contracts, not legal advice. Every document was generated for this dataset. The a

    azazelovesH100 · huggingface · 11h ago
  • Dataset
    keylazy/bab-slurpkeylazy · huggingface · 11h ago
  • Dataset
    Sssunset/RoboTrainingSssunset · huggingface · 11h ago
  • Dataset
    seharaman786/multilingual-model-blind-spots

    Blind Spots of Frontier Models: Roman Urdu and Code-Switching Research Question Does Qwen2.5-1.5B-Instruct consistently follow instructions and preserve language and scri

    seharaman786 · huggingface · 11h ago
  • Dataset
    mgwr74/Henrys_voice_datamgwr74 · huggingface · 11h ago
  • Dataset
    stefanocarrera/kic_D_train-java_Qwen3-4B_t0.5_g0_run0stefanocarrera · huggingface · 11h ago
  • Dataset
    phanerozoic/breakthrough-listen-gbt-waterhole

    Breakthrough Listen GBT L-band: HI and OH windows of nearby stars High-resolution spectrograms from the Breakthrough Listen open data archive, cut to the channels around

    phanerozoic · huggingface · 11h ago
  • Dataset
    canbingol/APIGen-MT-3k-turkish-translated

    APIGen-MT-3k-turkish-translated Bu veriseti, Salesforce/APIGen-MT-5k verisetinin Türkçe diline çevrilmiş ve sohbet şablonuna (chat template) uygun şekilde messages format

    canbingol · huggingface · 11h ago
  • Dataset
    lhz7891444/Annoy-PyEdu-Rs-Raw

    Annoy: This should be a paper Title 📑 Paper    |    🌐 Project Page    |    💾 Released Resources    |    📦

    lhz7891444 · huggingface · 11h ago
  • Dataset
    lhz7891444/Annoy-PyEdu-Rs

    Annoy: This should be a paper Title 📑 Paper    |    🌐 Project Page    |    💾 Released Resources    |    📦

    lhz7891444 · huggingface · 11h ago
  • Dataset
    yaseminbridges/mondo-evidence-datayaseminbridges · huggingface · 11h ago

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.