Hugging Face has released funes, a persistent memory layer for Claude Code, Codex, pi, and Hermes. It builds an index from existing local sessions and connects to an agent with a single command.
Why it matters: funes turns coding-agent session histories into searchable local memory and offers a way to share it across machines and agents.
A publicly available, low-cost recipe fine-tunes LFM2.5-350M with TRL's GRPO using about 500 samples and 100 training steps, raising the pass rate on IFStruct from 22.6% to 29.7%. Training can run on a free Colab or Kaggle GPU, while evaluation uses llama.cpp on a local MacBook. The code is open on GitHub.
A Hugging Face blogger reproduced Surya Narreddi's idea of getting a language model to paint watercolors using TRL and OpenEnv. The model writes about 150 lines of JavaScript with p5.brush to create the painting.
Microsoft has released GigaPath-Flash and GigaTIME-Flash, using a 22M-parameter ViT-S backbone distilled from the billion-parameter GigaPath encoder to sharply reduce the compute requirements of pathology foundation models. Both are open-source under Apache 2.0.
We use Inkling-Small to turn research abstracts into fast, useful summaries. Open weights × open science 🤝
Sentence Transformers v6.0 adds a fourth model type, MultiVectorEncoder, supporting ColBERT-style late-interaction retrieval and providing a complete training recipe.
METR research finds that AI's acceleration of science is uneven: cybersecurity-vulnerability reports sped up considerably from 2025 to 2026, mathematical research accelerated only slightly, and AI research itself showed no measurable acceleration. A university collaboration proposed SPADE, a self-play framework in which LLMs alternately generate executable training environments and solve them. Qwen3-30B-A3B scored 58.3 across the game-environment suite, an 8.1-point gain over its base model.
Microsoft Research has released Skala 1.1, a deep-learning DFT functional trained on 2.5 times as much data as the previous public version. It ranks first in 32 of the 55 GMTKN55 benchmark categories, with a weighted mean error of 2.8 kcal/mol. The release is available in CP2K and is being integrated into Psi4, FHI-aims, ORCA, and VASP. Microsoft has also introduced a living benchmark that tracks computational performance across versions.
Hugging Face's OlmoEarth Studio can now compute and export embeddings; its code and model weights are public. Through the UI or API, users can select a region, time range, encoder variant (Nano: 128 dimensions and 1.4M parameters; Tiny: 192 dimensions and 6.2M parameters; Base: 768 dimensions and 89M parameters), resolution, and imagery source, then export an int8 Cloud-Optimized GeoTIFF.
A new Hugging Face paper proposes two systems improvements that sharply reduce the training cost of distilling LLMs: cache the teacher model's top-100 logits offline so teacher and student do not need to reside in GPU memory simultaneously, and use a fused, chunked KL loss to avoid generating a full vocabulary-by-sequence-length matrix.
DharmaOCR scored 0.925 on a Portuguese OCR benchmark, ahead of Mistral OCR4 at 0.798 and Unlimited-OCR at 0.7587. Two-stage training with Portuguese supervised fine-tuning followed by DPO concentrates all model parameters on Brazilian Portuguese, giving it a domain-specific advantage in extraction quality and stability.
Hugging Face has introduced a “kernel” repository type on the Hub for the Kernels project, along with two security measures: trusted publishers and code signing. By default, only kernels from trusted publishers are loaded; other sources require explicitly passing `trust_remote_code=True`.
Hugging Face has released ScarfBench, an open-source benchmark for evaluating AI agents on enterprise Java framework migrations. It covers the Spring, Jakarta EE, and Quarkus ecosystems, with 34 applications, 102 framework implementations, and 204 migration tasks.
A Hugging Face blog introduces DiScoFormer (Density and Score Transformer), a single model that estimates both the density and score of a data distribution in one forward pass, without retraining for a new distribution.
The alignment team at the UK's AI Security Institute and Timaeus have jointly founded Sequent, a nonprofit research organization. It argues that “alignment is not on track” and plans to grow to 40–80 full-time staff within two years, with an initial fundraising target of $100–150 million. Research areas include scalable oversight, learning theory, heuristic argumentation, game theory, and personas.
After acquiring Emmi AI, Mistral is building foundational physical AI for aerospace, automotive, semiconductors, and energy. Its announced results include AB-UPT, which processes raw geometry with 9M surfaces and 140M volume cells on a single GPU without remeshing, and NeuralDEM, an end-to-end deep-learning surrogate for large multiphysics processes.
Import AI 457 focuses on three studies. SentinelOne analyzed fast16.sys, a virus from more than 20 years ago that corrupted results from high-precision software such as LS-DYNA 970, PKPM, and MOHID by tampering with in-memory calculations, sabotaging scientific and engineering work.
Google Research has proposed MoGen (Neuronal Morphology Generation), a model that generates synthetic neuron shapes with point-cloud flow matching. The data supplements training for the PATHFINDER reconstruction model, reducing mouse-axon reconstruction error by 4.4%, mainly by decreasing merge errors. At whole-mouse-brain scale, the work is equivalent to saving 157 person-years of manual proofreading. MoGen is available as an open-source model.
