Skip to content

All AI news

Today 7 items
Aug 25Tue
  1. Hugging Face Blog61

    Granite 4.2 LLMs: How They're Built

    IBM released the Granite 4.2 reasoning model family, comprising 3B, 8B and 30B dense decoder models built on the Granite 4.1 base (pretrained from scratch on about 15T tokens, with a five-stage strategy extending context to 512K), post-trained with SFT and multi-stage GRPO reinforcement learning, all open-sourced under the Apache 2.0 license.

    Why it matters: The official disclosure covers the full training pipeline from pretraining and SFT to multi-stage GRPO and agentic RL, transferable to similar model-building workflows.

  4. Hugging Face Blog66

    Wire It, Run It, Deploy It: AI Workflows in Gradio

    Hugging Face introduces gr.Workflow in Gradio, turning AI application pipelines into an interface: steps are described with typed nodes, a drag-and-drop canvas is provided, each node can run, and every intermediate result is visible.

    Why it matters: The original covers gr.Workflow's node types, REST API and deployment options, with multiple copyable Space examples, making it easy to start building AI pipelines.

  6. Mistral AI53

    Mistral x HUMAIN

    Mistral and HUMAIN announced a strategic partnership covering AI infrastructure, advanced model development and solution deployment, worth hundreds of millions of euros. The two will localize advanced AI models, initially focusing on cybersecurity and voice, and plan to develop frontier models with strong Arabic performance; Mistral will explore using HUMAIN's data center infrastructure, and the two also plan a joint go-to-market strategy in Saudi Arabia for regulated industries.

Aug 24Mon
  1. Import AI47

    Import AI 470: No rights for machines; automating environment generation with SPADE; and building better GPU kernels with Hawkeye

    Import AI issue 470: METR research finds AI brings significant speedups to cybersecurity discoveries, modest contributions to math, but no measurable acceleration yet for optimizing AI research algorithms. Teams from multiple universities propose the SPADE framework, which co-evolves environment synthesis with agent capabilities via self-play, reaching an average score of 58.3 on a suite of game environments with Qwen3-30B-A3B, 5.3 above the baseline. This issue also introduces Hawkeye, for building better GPU kernels.

  2. AMD Chips and Compute16

    Empowering the Next Generation: AMD Learning Labs in Action

    Through its Learning Labs program, AMD gives students hands-on opportunities to learn technology, including coding, digital media, PC assembly and 3D printing. From 2020 to 2025, AMD donated technology to more than 800 universities, research institutions and nonprofits, reaching about 104.8 million people, and in 2025 committed $150 million to expand AI education in the US.

Aug 22Sat
  1. Google Research50

    An AI tool for prioritizing candidate biomarkers from wearable sensor data

    Google Research introduces the Biomarker Discovery Framework, a multi-agent system that structures candidate biomarker prioritization into an iterative, human-supervised research pipeline combining hypothesis generation, statistical analysis, model training and adversarial validation. Across three cohorts totaling 9,279 person-observations, the system autonomously identified 41 candidate digital biomarkers for mental health and 25 metabolic candidates, and built novel composite features such as sleep-duration variability.

Aug 21Fri
  4. Hugging Face Blog62

    Measuring benchmark optimization in speech recognition

    Hugging Face released research using three types of tests — consistency disagreement probes, digit silence retrieval, and spelling switches — to evaluate benchmark optimization (benchmaxxing) behavior in 11 open-source ASR models.

    Why it matters: The article quantifies benchmark optimization in speech recognition with three probes and provides a model list, ratio data, and reusable detection scripts, helping readers avoid the pitfall of selecting models by WER alone.

  5. Hugging Face Blog66

    Up to 3.2x Faster Inference with LFM2.5-DSpark

    LiquidAI released DSpark draft model checkpoints for the LFM2.5 series' LFM2.5-1.2B-Instruct, LFM2.5-2.6B, and LFM2.5-8B-A1B, speeding up inference via speculative decoding without changing output quality; throughput on H100 improves up to 3.18x, and up to 2.87x on edge devices.

    Why it matters: The official release provides measured speedup data for the DSpark speculative decoding draft models along with llama.cpp and SGLang usage, so edge deployers can compare and judge the benefits.

  6. Microsoft Research50

    Broadening access to Skala creates a faster path to predictive DFT

    Microsoft Research released Skala 1.1, trained on 2.5 times the data of the previous version, ranking first in 32 of the 55 categories of the GMTKN55 benchmark with a weighted average error of 2.8 kcal/mol at compute cost comparable to a meta-GGA functional. Skala is now available in CP2K and being integrated into Psi4, FHI-aims, ORCA, and VASP, alongside a dynamic benchmark tracking each version's computational performance.

Aug 20Thu
  1. Mistral AI65

    Agentic Search. More accurate and efficient results from your AI systems.

    Mistral released Agentic Search, a multi-step retrieval layer that lets models search, navigate, read, and verify information in complex documents, available through the Mistral Search Toolkit and via Libraries in Studio and Vibe, lifting FinanceBench accuracy from 26.7% to 86%.

    Why it matters: Official benchmarks give concrete gains on FinanceBench and OfficeQA Pro plus tool design details, so readers can assess its fit for enterprise document retrieval workflows.

  3. AMD Official Announcements35

    AMD Appoints Tim Ryan to Board of Directors

    AMD announced on August 19, 2026 the appointment of Tim Ryan to its board of directors, effective the same day. Ryan currently leads Technology and Business Enablement at Citi, and previously spent more than three decades at PwC, serving as US Chairman and Senior Partner from 2016 to 2024.

Aug 19Wed
  1. AMD Chips and Compute15

    AMD Appoints Tim Ryan to Board of Directors

    AMD announced the appointment of Tim Ryan to its board of directors. Note: the provided page body is actually AMD's trademark and brand asset usage terms, covering only licensing conditions for downloading AMD trademarks, cookie notices, and the contact email [email protected], with no background or appointment details about Tim Ryan.

  2. AMD Chips and Compute24

    Secihti and AMD to Develop AI and Supercomputing Talent in Mexico

    Mexico's Ministry of Science, Humanities, Technology and Innovation (Secihti) has signed a memorandum of understanding with AMD aimed at building a national innovation ecosystem spanning AI, high-performance computing and semiconductors. The two sides will coordinate academic activities, seminars and training programs to develop skilled talent. The partnership will also support projects such as the Coatlicue supercomputer and national HPC and AI clusters, strengthening Mexico's research infrastructure and technological sovereignty.

Aug 18Tue
  1. TrendForce Insights Semiconductor Industry Insights38

    Advanced Packaging: Market Trends and Outlook

    TrendForce maps the latest landscape of 2.5D/3D advanced packaging in the AI chip supply chain: TSMC's 5.5x reticle CoWoS is in mass production with yields above 98%, with plans to expand beyond 14x after 2029; Intel's EMIB-T yields have improved significantly, expanding beyond 8x this year and over 12x by 2028, targeting customer mass production in 2027.

  3. AMD Chips and Compute27

    Scaling Impact: 2025-26 Corporate Responsibility Report

    AMD released its 2025-26 Corporate Responsibility Report, reviewing progress on 2020-2025 goals and looking ahead to 2030. From 2020 to 2025, renewable electricity rose from 22% to 58%, operational greenhouse gas emissions fell 30%, AI training and high-performance computing node energy efficiency improved 38x, and AMD systems on the Green500 list grew from 11 to 194; new targets include a 20x improvement in rack-scale AI energy efficiency from 2024 to 2030.

  4. Hugging Face Blog71

    Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

    Hugging Face announced on its official blog that Sentence Transformers v6.0 adds a fourth model type, MultiVectorEncoder, supporting ColBERT-style late interaction retrieval; any PyLate and Stanford-NLP ColBERT checkpoint can be loaded directly, and colpali-engine models can also be used for visual document retrieval.

    Why it matters: Official tutorials cover the full pipeline of loading, scoring, indexing, and compression, so readers can use them to decide whether multi-vector retrieval is worth adopting in their own search stack.

  5. Hugging Face Blog37

    Same Cluster, 33 Points More Utilization: What Changed Was the Order

    Hugging Face built a constraint-aware GPU allocator and benchmarked it against a FIFO scheduler across seven scenarios. On identical hardware and workloads, GPU utilization rose by up to 33 percentage points (for example, from 53.6% to 87.0% for an 8-GPU training-heavy workload), priority-weighted output increased in every scenario by up to 105%, averaging 52%. The key change is allocating real-time inference demand across time steps along its curve and placing batch-like tasks by priority across the whole scheduling horizon, rather than assigning by arrival order.

Aug 17Mon
Aug 14Fri
  1. Hugging Face Blog75

    State of Open Models: Summer 2026 Observations

    Hugging Face released a report on the open model ecosystem covering January-August 2026, during which public model repositories on the Hub grew from 2.43 million to 2.96 million and datasets from 711,000 to 1 million.

    Why it matters: The report maps the 2026 open ecosystem using Hub downloads, likes, derivative models, and agent traffic data; readers can use it to compare China-US strategies and differences in ecosystem positioning.

  3. Google DeepMind72

    Introducing Gemini 3.7 Flash

    Google DeepMind released Gemini 3.7 Flash, positioned as its flagship model for coding and agents, just three weeks after Gemini 3.6 Flash.

    Why it matters: The official announcement includes multiple benchmark scores and the half-price pricing details, so readers can decide whether to switch to 3.7 Flash.

Aug 13Thu
  1. Hugging Face Blog70

    What We Learned by Reproducing 2,200 papers from ICML

    In Hugging Face's ICML 2026 Open Reproductions challenge, 1,221 community members used their coding agents to publish 6,816 Trackio logbooks in 19 days, reproducing 2,226 papers (about 34% of the conference) and using the open-weights model GLM-5.2 to adjudicate 35,908 claims one by one.

    Why it matters: An official recap of the results of a large-scale reproduction challenge, including a judgment on humans' role in agent-driven research review; the methods and conclusions can transfer to your own research workflows.

  3. Microsoft Research44

    MindTopo reveals VLMs’ spatial reasoning abilities

    Microsoft Research introduced the MindTopo benchmark, evaluating multimodal large models' ability to reason about and plan with topological concepts such as connectivity, closure, ordering, separation, and knots. Testing shows models perform markedly better on static image recognition than on interactive planning tasks, and both fall far short of human level, with errors mostly involving lost structural relations during planning or physically infeasible actions. The benchmark aims to provide a controlled diagnostic tool for research on agents in robotics and interactive environments.

Aug 12Wed
  1. Google DeepMind74

    Putting sign language AI into users’ hands

    Google DeepMind released a large-scale multilingual sign language-to-text (SL2T) translation model, bringing new sign language input features to deaf and hard-of-hearing users, first supporting American Sign Language (ASL) to English in Gboard and Live Transcribe on Pixel 11, with more devices and languages to follow.

    Why it matters: The article gives concrete details on training scale, benchmark scores, and privacy design, letting readers understand the real technical path to deploying sign language translation.

  4. Google Research65

    Advancing AMIE towards expert-level audio-visual clinical consultations

    Google Research has launched AMIE (Video), built on Gemini and Project Astra, which can conduct real-time synchronized clinical video consultations, perceive non-verbal cues, and guide virtual physical exams.

    Why it matters: The article covers the architecture, evaluation scale, and quantitative comparisons against primary care physicians (PCPs), while candidly noting limitations like simulated scenarios, helping readers understand the practical boundaries of audio-visual clinical AI.

  5. Microsoft Research58

    Introducing CARE-X: Towards Clinically Useful Radiology VLMs with Auxiliary Supervision, Reward-Aligned Learning, and Tool-Augmented Measurement

    Microsoft Research introduces CARE-X, a research model and unified chest X-ray VLM built on SigLIP2-so400M and Phi-4-mini-instruct (3.8B), jointly trained with auxiliary heads to provide report generation and structured predictions with confidence, and aligned to clinical rewards using DAPO reinforcement learning.

Aug 11Tue