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

Everything Launches Models Datasets Papers Tools and code News

Live now

Connecting…

Just launched

ReportedAuroraTTS-100K Releases1h ago ReportedLiquid AI Launches RIPL, a Real-Time Inference Engine for Sub-50 ms LFMs4h ago ReportedClaude Duet4h ago ReportedDeepSeek and MiniMax Clear China’s AI Export Rules for Overseas Open-Source Releases4h ago ReportedGradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.04h ago ReportedRerank 55h ago ReportedBenchmark: Recursive's Flint 40B A2.5B Q4 achieves 128 tok/s on MacBook Pro M4 Pro 48GB using Kapitsa, an open-source Apple Silicon optimized engine, Claude/Codex/Pi/OpenCode compatible.9h ago ReportedClaude Duet AI agent10h ago

Today

Dataset
rainbowrobotics/JIMTOF_newrobot_20261022_085505rainbowrobotics · huggingface · 10h ago
Model
HimanshuPathak/omnivoice-hindi-september-intermediate-archive-20261022-gatedHimanshuPathak · huggingface · 10h ago
Dataset
flowyrepo/herdsman_versionsflowyrepo · huggingface · 10h ago
Dataset
flowyrepo/local-model-runtimesflowyrepo · huggingface · 10h ago
Model
Betswish/seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B-Instruct_en-de_LIDR_5e-06Betswish · huggingface · 10h ago
Model
qijiashen/agentrl-oe-task2-20261022-110003qijiashen · huggingface · 10h ago
Model
okingjo/Ronova 若娜瓦
Brief Lora model of Ronova from Genshin Impact. Trained with 1024x1024 resolution for SD1.5, 1536x1536 for SDXL, 2048x2048 for pony and Illustrious version. The training
okingjo · civitai · 10h ago
Model
RDMPrompt/Kagerou Imaizumi (Touhou Project) - 今泉影狼 (東方)
Kagerou Imaizumi from Touhou Project Anima V1 Kagerou Imaizumi : imaizumi kagerou, 1girl, solo, brown eyes, brown hair, long hair, wolf girl, wolf ears, wolf tail, sharp
RDMPrompt · civitai · 10h ago
Model
andraste/胡桃のあ Kurumi Noah 2.0 [40 Outfits] / ぶいすぽっ！ Vspo! [Illustrious]
1.0と2.0のプロンプトは一部異なります。詳細はそれぞれの説明を参照してください。 1.0与2.0的提示词有一部分不同，具体请参考各自说明。 The prompts for 1.0 and 2.0 are partially different. Please refer to their respective descriptions
andraste · civitai · 10h ago
Model
Betswish/seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B-Instruct_en-fr_LIDR_5e-06Betswish · huggingface · 10h ago
Model
Quiho/Prefectious_XL_NSFW_v1.0_checkpointQuiho · huggingface · 10h ago
Dataset
EdgeArbiter/2LEVEL_DataEdgeArbiter · huggingface · 10h ago
Model
Betswish/seed0_sample24000_xcsqa_google-gemma-3-4b-it_multi_allpairs_DCO_1.0Betswish · huggingface · 10h ago
Model
benyunzhao/umi-arena-pi05-strict3h-fft-20000benyunzhao · huggingface · 10h ago
Model
Betswish/seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B-Instruct_en-ar_LIDR_5e-06Betswish · huggingface · 10h ago
Dataset
emmilina/fatimahemmilina · huggingface · 10h ago
Model
BSC-NLP4BIA/GreekMedNLP-eurobert-210m-2025-2026BSC-NLP4BIA · huggingface · 10h ago
Dataset
eunzee/PCEvaleunzee · huggingface · 10h ago
Model
Kushagra0011/whisper-large-v3-hinglish-loraKushagra0011 · huggingface · 10h ago
Model
Betswish/seed0_xcsqa_google-gemma-3-4b-it_multi_0.1_LIDR_5e-06Betswish · huggingface · 10h ago
Model
Aratako/Drama-3-JapaneseAratako · huggingface · 10h ago
Dataset
ShaswatRAJ/rppg-deepfake-detectionShaswatRAJ · huggingface · 10h ago
Dataset
PureOne/EVIE-STREAM
EVIE-STREAM Training High-Fidelity Digital Clones on Noisy, Fragmented, Interrupted, and High-Latency Data While Preserving Maximum Behavioral Accuracy and Long-Term Iden
PureOne · huggingface · 10h ago
Model
Betswish/seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B-Instruct_en-es_LIDR_5e-06Betswish · huggingface · 10h ago
Dataset
Ksri12693/FreshRetailNet-50K
FreshRetailNet-50K Dataset Overview FreshRetailNet-50K is the first large-scale benchmark for censored demand estimation in the fresh retail domain, incorporating approxi
Ksri12693 · huggingface · 10h ago
Model
BSC-NLP4BIA/GreekMedNLP-bert-base-greek-uncased-v1-2025-2026BSC-NLP4BIA · huggingface · 10h ago
Model
gabe-adib/ove3-k3align-refit-c1e18-h384-lrx0.35gabe-adib · huggingface · 10h ago
Dataset
root-sys-n/aurora-mini-diffusion-all
Aurora Mini Diffusion - All Generations All images generated by AuroraIMG-6M from 10,000 diverse prompts x 20 seeds. Carries English and Turkish captions. Statistics Fiel
root-sys-n · huggingface · 10h ago
Model
aaronr43210/terrapulse-qwen3-1.7b-GGUF-v2aaronr43210 · huggingface · 10h ago
Model
aaronr43210/terrapulse-qwen3-1.7b-lora-v2aaronr43210 · huggingface · 10h ago
Model
aaronr43210/terrapulse-qwen3-1.7b-v2aaronr43210 · huggingface · 10h ago
Dataset
kaiseong/rainbowroboticskaiseong · huggingface · 10h ago
Model
mradermacher/granite-speech-5.0-2b-GGUFmradermacher · huggingface · 10h ago
Dataset
Mobiusi/Graduate-Entrance-Exam-Mathematics-Multi-Step-Question-Answering-Dataset
Graduate Entrance Exam Mathematics Multi-Step Question Answering Dataset This dataset contains graduate entrance exam mathematics questions paired with reference solution
Mobiusi · huggingface · 10h ago
Model
gabe-adib/ove3-k3align-refit-c1e18-h256-lrx0.35gabe-adib · huggingface · 10h ago
Model
ZhaoyangWei/sdft-weightsZhaoyangWei · huggingface · 10h ago
Dataset
Mobiusi/Material-Testing-Instrument-Accessory-Compatibility-Relations-Dataset
Material Testing Instrument Accessory Compatibility Relations Dataset This dataset covers equipment compatibility, specimen suitability, and intended-use information draw
Mobiusi · huggingface · 10h ago
Dataset
khtsly/g1khtsly · huggingface · 10h ago
Model
Ryanham1lton/SwadloonRyanham1lton · huggingface · 10h ago
Model
mradermacher/VanMach-7B-GGUFmradermacher · huggingface · 10h ago

Newer Older

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.