AI BriefingHuggingFace Blog news
The latest AI and developer news from HuggingFace Blog, with short summaries.
Visit website →IBM releases Granite Switch 4.1.1 with updated Speech ASR and Decision adapters
The patch release updates the Granite Speech ASR model for 10 languages and improves the Decision adapter without recomputing the base model.
Hugging Face and CMU Launch Project ProtoStar to Make Publicly Funded Science AI-Ready
The initiative partners with universities and industry to release scientific artifacts in machine-actionable formats, starting with a CMU-hosted climate prediction challenge on November 2.
Mistral Large 4 Released on Hugging Face with 1M Context and Top Cybersecurity Scores
The 1T parameter model features a 1M token context window and ranks 5th on the Artificial Analysis Intelligence index for open-weight models.
Hugging Face Releases AnySpeechRecognizer and AnyDecisionModel Swift Packages for Apple Platforms
The packages wrap Apple's new SpeechAnalyzer API with NVIDIA and Alibaba models, achieving 1.55% WER on LibriSpeech test-clean, and enable typed decision outputs via six backends including MLX and llama.cpp.
NVIDIA Releases Nemotron 3.5 Transcribe and Granary v2 for Multilingual Speech Processing
Open-weight model supports 40 languages and speaker diarization, with a 30B parameter architecture using 4B active parameters per token.
Baidu Unlimited-OCR 2: Parsing hundreds of pages by maintaining context across pages
Baidu has released Unlimited-OCR 2, which uses cross-page context to parse long documents.
Ai2 releases benchmark showing LLM novelty judges are unreliable on AI-generated research ideas
A new benchmark from Ai2 reveals that small prompt changes can shift LLM novelty judgment accuracy by over 50 percentage points, and simple baselines outperform expensive purpose-built systems.
Ai2 and University of Pennsylvania develop LitPivot for literature-informed research ideation
In a study with 17 researchers, domain experts rated ideas developed with LitPivot as better grounded in literature than those developed with a text editor and AI chat tool.
Hugging Face Launches TorchNeuron Model Support Space for AWS Trainium Compatibility
The new Space verifies if popular Hugging Face models run unmodified on Trainium2, providing execution scripts and a public dataset of results.
PickMicrosoft Releases ThinkingBox Training Stack for RL Post-Training of Agents Using Database State Rewards
The method uses executable database state checks as binary rewards to fine-tune agents, reducing the cost per dependable task from $24.36 to $8.00 for Qwen3.8-27B.
PickAi2 Releases OlmoEarth v1.3 with State-of-the-Art Global Satellite Embeddings
The model generates 128-dimensional vectors for every 10-meter pixel on Earth, outperforming AlphaEarth Foundations and Tessera v2 across six continents.
IBM Research releases Granite Timeseries Ensemble for zero-fine-tuning probabilistic forecasting
The ensemble achieves a MASE of 0.683 on GIFT-Eval, outperforming its best individual member without requiring fine-tuning.
NVIDIA Kumo Relational Outperforms Supervised ML on Relational Data Without Task-Specific Training
The pre-trained foundation model requires no feature engineering or table flattening, using up to 10,000 historical examples for single-pass inference on databases with over 500 billion rows.
NVIDIA and Pollen Robotics Demonstrate Microduck Robot Locomotion Training in Isaac Lab 3.0
NVIDIA, Pollen Robotics, and Hugging Face introduce a workflow to train walking policies for the Microduck robot in Isaac Lab 3.0 and deploy them via ONNX.
VocabX Method Cuts SLM Inference Cost by 77% via Cost-Weighted Vocabulary Specialization
A new method specializes Small Language Model vocabularies to reduce measured inference costs by 77% while improving quality scores on domain-specific tasks.
TorchNeuron Profiling on AWS Trainium: System-Level Analysis and Fusion Benefits
A Hugging Face blog post details system-level profiling methods for TorchNeuron on AWS Trainium, showing that operator fusion reduces RMSNorm device time from 2,355 µs to 375 µs.
PickIBM releases Granite Speech 5.0 with improved ASR accuracy and 1.6x faster Turbo CTC
New 2B model ranks #2 on Open ASR Leaderboard while updated Turbo CTC achieves 19,870 RTFx.
NVIDIA Releases Kumo Forecast, a Time-Series Model Combining Pretrained Representations with Historical Retrieval
The open-source model uses a MOMENT-1-large backbone and DARR mode to blend direct predictions with k-nearest-neighbor retrieval from historical data.
NVIDIA and University of Washington Release RFD4-Proteina on Hugging Face Spaces
The unified flow-matching model outperforms Proteina-Complexa by 5.8x in design yield while integrating sequence and all-atom structure generation.
PickNVIDIA Releases SimReady Central: Open Library of 2,000+ Simulation-Ready 3D Assets on Hugging Face
The library provides pre-configured OpenUSD models with physics and thermal metadata to accelerate digital twin construction and robotics training.