Villanova2: Italy's open-weight multilingual MoE family, built for Europe
Villanova.AI released Villanova2, an open-data-and-weights family of language and vision-language models trained in Italy: a 15.5B-total / 3.9B-active MoE and a compact 2.4B dense model, each in base, instruct, and vision-language versions.
Why it matters
A European open-weight family tuned for five European languages with a document-understanding focus: the 15B-A3B scores within 0.4 points of Qwen3-8B on the team's 85-benchmark suite, with GGUF builds for on-device use.
What you could build with it
A small legal-tech team in Milan could run the 2B vision-language model on a local workstation to extract structured data from Italian contracts and invoices, keeping client documents on-premises instead of sending them to a cloud API.
Does it hold up?
Too early to judge — released 11 Nov with weights on Hugging Face and a technical report still to come; no independent benchmarks or deployment reports exist yet.
Built with Villanova2
- Villanova2: Made in Italy, built for Europearticle · Announcement with the full specs: 15.5B-A3B MoE and 2.4B dense backbones, 4.7T pretraining tokens, 131K context, five languages, GGUF builds for llama.cpp.
- VillanovaAI on Hugging Facehf · Org page hosting the Villanova2 model family, including the 2B and 2B-VL checkpoints and GGUF builds.
Learn more
First spotted on hf: source.
More AI models
VoiceStudio: open-source, fully-local ElevenLabs alternative for voice cloning, dubbing, and audiobooksdebpalash/VoiceStudio is a desktop app that runs voice cloning, voice design, video dubbing, dictation, transcription…models · JEV 0.75Edge0-145B: open 145B MoE that runs on a single consumer GPUEdge0 open-sourced 145B and 120B latent-MoE models derived from Qwen3.8-Flash-Next, plus an inference framework that…models · JEV 0.74Strata: run a 125B model on a gaming PCFree, MIT-licensed local inference engine that runs Qwen3.8-Flash-Next (125B MoE, about 6B active per token) on…models · JEV 0.73Google AI Edge Gallery: offline speech recognition and synthesis on-deviceGoogle added fully offline speech recognition and text-to-speech to AI Edge Gallery, with ML Kit APIs on Android and…models · JEV 0.73Soup: fine-tune an 8B LLM on a 4 GB laptop GPU from one YAML fileAn open-source CLI that fine-tunes LLMs from a single YAML file, with layer streaming that trains an 8B model on a 4…models · JEV 0.73Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.