Hugging Face released Grabette, an open-source system recording manipulation demonstrations with a handheld gripper and two cameras, producing robot-ready datasets without robots or teleoperation equipment. The handheld hardware costs about €490 in materials, with the accompanying motorised Gripette gripper around €120. Hardware CAD, Raspberry Pi collection software and browser-based processing are all open-source.
DharmaOCR scored 0.925 on a Portuguese OCR benchmark, ahead of Mistral OCR4's 0.798 and Unlimited-OCR's 0.7587. It specialised through two-stage training: supervised fine-tuning on Portuguese corpora, then DPO to stabilise inference. The author argues that concentrating parameters on one language remains a structural advantage despite emerging architectures.
Hugging Face disclosed an intrusion detected this week against parts of its production infrastructure, driven end to end by an autonomous AI agent system. Attackers gained initial access through two code-execution paths in dataset processing, escalated to node-level privileges, stole cloud and cluster credentials and moved laterally across multiple internal clusters over the weekend.
Hume released Real World VoiceEQ, a speech evaluation benchmark covering over 40 proprietary and open-source voice models, more than 15 evaluation dimensions and over 60 metrics across ASR, TTS, S2S and speech understanding.
Thinking Machines released Inkling on Hugging Face, a multimodal MoE model with around one trillion parameters, a one-million-token context and native image, text and audio inputs, alongside Inkling-Small with 276 billion total and 12 billion active parameters.
The third Hugging Face PyTorch profiling instalment traces attention implementation with a profiler. Comparing naive attention with in-place masking shows replacing masked_fill with masked_fill_ removes one GPU Memcpy kernel per forward pass.
Hugging Face announced that vLLM’s transformers modelling backend now matches or exceeds the throughput of vLLM’s handwritten native implementations across several LLM architectures. It uses torch.fx for static graph analysis and ast to rewrite source code, dynamically applying inference-related layer fusion at runtime to match custom-code performance.
Hugging Face and Amazon SageMaker AI introduced deep-link integration. Clicking Customize on SageMaker AI or Deploy on SageMaker AI on supported model pages opens SageMaker Studio with the model preloaded and environment configured automatically.
At Build 2026, Microsoft announced Foundry Managed Compute and a Hugging Face model collection, with weekly updates to open-weight models and one-click deployment to Foundry's managed GPU platform.
Hugging Face released LeRobot v0.6.0, introducing world model policies VLA-JEPA, FastWAM and LingBot-VA, new VLAs GR00T N1.7, MolmoAct2, EO-1, EVO1 and Multitask DiT, and a unified reward model API with Robometer and TOPReward for the robotics learning loop.
Hugging Face and SkyPilot released an integration allowing a Hugging Face Bucket or any model, dataset or Space repository to be mounted into SkyPilot tasks with an hf:// URL and an existing HF_TOKEN, running across more than 20 clouds, Kubernetes, Slurm and local environments.
Hugging Face's fourth PRX instalment details its data strategy: mixing public and internal pretraining datasets, regenerating long image captions with a VLM and converting them for streaming training. It uses Lance for construction and filtering and MDS for streaming. After switching to Qwen3-VL, text latents are computed during training, with measured throughput loss of around 3–4%, or roughly one extra day for 30 days of training.
Hugging Face introduced a new kernel repository type for Kernels, along with trusted publishers and code signing. By default, only trusted publishers' kernels load; other sources require explicit trust_remote_code=True.
Mistral released Leanstral 1.5 under Apache-2.0, with 119B total and 6B active parameters and major formal verification improvements. It achieves 100% on miniF2F, solves 587 of 672 PutnamBench problems, and sets current best results of 87% on FATE-H and 34% on FATE-X.
Hugging Face and Cerebras jointly demonstrated a real-time speech-to-speech pipeline, accelerating Gemma 4 31B inference with Cerebras and combining Nvidia Parakeet speech recognition with Alibaba Qwen3TTS synthesis.
Hugging Face released ScarfBench (Self-Contained Application Refactoring Benchmark) to evaluate AI agents migrating enterprise Java applications between Spring, Jakarta EE and Quarkus.
Google Research released TabFM for tabular classification and regression. It reframes table prediction as in-context learning, producing predictions in one forward pass without manual training, hyperparameter tuning or feature engineering.
Every Eval Ever (EEE) and Hugging Face Community Evals are now interoperable, allowing cross-publication of evaluations with links to complete records. EEE stores around 229,000 results across more than 22,000 models and 2,200 benchmarks from 31 reporting formats.
Hugging Face proposed DiScoFormer (Density and Score Transformer), estimating a distribution's density and score simultaneously in one forward pass from a set of data points, without retraining for new distributions.
Google DeepMind released experimental open-source model DiffusionGemma, generating text blocks in parallel through text diffusion and achieving up to 4 times faster inference on dedicated GPUs.
Mistral AI released its first text-to-speech model, Voxtral TTS, with 4B parameters and support for nine languages: English, French, German, Spanish, Dutch, Portuguese, Italian, Hindi and Arabic. It is available through the API and Mistral Studio at $0.016 per 1k characters.
Mistral released Voxtral Transcribe 2, comprising Voxtral Mini Transcribe V2 for batch transcription and Voxtral Realtime for real-time use.
Mistral AI released Voxtral speech-understanding models in 24B and 3B versions, both under Apache 2.0 and available via its API. With a 32k-token context, they handle up to 30 minutes of transcription or 40 minutes of audio understanding, including built-in question answering and summaries, automatic multilingual detection and voice-triggered function calls, inheriting Mistral Small 3.1's text capabilities.
Mistral AI released its first reasoning model, Magistral, comprising the 24B open-source Magistral Small and enterprise-focused Magistral Medium.
Mistral AI and All Hands AI jointly released Devstral, an agentic LLM for software engineering tasks, open-source under Apache 2.0.
Mistral AI released Mistral Small 3.1, improving text performance and multimodal understanding over Small 3, expanding context to 128k tokens and reaching 150 tokens per second, under Apache 2.0.