#Data/Training

AILlamaIndex announced OpenDocRouter, a single API that routes document parsing requests to any of 10 frontier and open-source models at launch, including Claude Opus 5.5, Gemini 3.8 Flash, GPT-6 Luna, MinerU2.5-Pro, and PaddleOCR-VL-1.6. Users can switch models in one line with the same request and markdown output, and each model is scored on ParseBench for quality and cost. Pricing is per-token, failed pages are not charged, and the service costs $0.86 to $48.82 per 1,000 pages depending on the model.

AIMicrosoft Research Asia has open-sourced Agent Lightning v1.0, a roughly 3,500-line agentic RL framework that trains the same agent harness used in deployment. In an end-to-end coding agent pipeline, Qwen3.5-9B rose from 41.8% to 56.4% Pass@1 on SWE-bench Verified using about 6,000 training samples. The framework runs agents as standard Kubernetes jobs without paid commercial sandbox services.

Why it matters: The source shows how training with the deployed agent harness avoids rebuilding agents, and reports concrete SWE-bench Verified gains from about 6,000 samples.

AIAi2 is making new Stage 1 checkpoints available, with the byte-level components already trained and the original model's weights left unchanged. Researchers can use them to test new architectures and train the full system without repeating that initial stage.

AIAi2 introduces Bolmo, a model that operates on bytes rather than tokens, which the team hopes will lead to larger models adapting how they represent information across languages and domains. Because bytes can also represent images and audio, the approach could eventually extend beyond text.

AIAi2 converted Qwen 3 8B and Llama 3 8B into byte-level models named Bwen 8B and Blama 8B. Both come close to matching their source models' performance in Ai2's evaluations, and Bwen 8B outperforms Bolmo 7B across its aggregate evaluation suite.

AIAi2 describes a recipe that adapts existing models to process raw bytes through a brief additional training run. The approach keeps the model's core intact and adds components that group bytes into variable-length patches for processing, then expand outputs back to byte-level predictions of the next byte.

AIMost language models split text into subwords from a fixed vocabulary, which can obscure spelling details and fragment meaningful units in code or math. Byte-level models instead work directly with the bytes computers use to represent text.

AIAi2's paper on retrofitting language models to operate over bytes, the approach behind Bolmo, has been accepted to Nature. The team is also releasing new checkpoints that extend the method from Olmo to Qwen and Llama.

AINVIDIA's cuOpt decision optimization software now targets problems with 100 million variables using mPDLP. The source says cuOpt already delivers more than 10x speedups over CPU solvers, while detailing little else about mPDLP's design or availability.

AIGoogle released the Google AI Edge Foresight app for macOS, powered by Gemma 4 and EmbeddingGemma 2. The app can connect to Google Drive to build a knowledge graph and, when transcription is active, uses local Gemma 4 E4B or Gemma 4 12B models to transcribe voice notes into new documents. EmbeddingGemma 2 is an open-weight, Apache 2 licensed 740M-parameter multimodal embedding model with an 8K context window.

AINVIDIA's VERA turns benchmark trajectories into over 9,000 restartable sandboxes with rubric scoring and updates both model weights and the agent harness together. A harness edit is kept only if it adds at least 5 points on the development set, and a checkpoint is rejected if its score drops more than 20%. At 27B, the co-evolved agent scores 71.6 on AutoCoWorkBench, above Claude Opus 4.8, and the environment corpus is open-sourced.

AINVIDIA's cuPhoton targets the computational bottleneck in scientific image pipelines, where data from observatories, telescopes, lasers, and X-ray light sources arrives faster than CPU-bound processing can handle. The source says the bottleneck is usually the whole path from raw sensor data to decision, not one slow kernel. The available text does not give benchmark figures, pricing, or availability details.

AIGoogle's SynthID portal at synthid.com lets users upload an image, video, or audio file to check it. The portal scans the file for a SynthID watermark from Google or its partners, which indicates AI generation.

AIGoogle DeepMind has made SynthID Detector publicly available, letting anyone check whether online content was generated using Google AI or tools from partners including OpenAI, NVIDIA, and Kakao. Apple is listed as coming soon. The tool is accessible at synthid.com.

AIGoogle says its SynthID detector now supports watermarks from all of its partners and is available to anyone on a new SynthID.com website. Previously, outside users could only check for Google's own watermarks by asking Gemini, and the detector did not flag images from other providers such as ChatGPT.

AIHiPHI is a 617.5-hour whole-body human motion dataset captured with optical motion capture at sub-millimeter accuracy, including 245.7 hours of human-object interaction with synchronized object trajectories and meshes. The dataset organizes coverage using FrameNet, a linguistic framework for human action. The white paper also reports results from policies trained on HiPHI and deployed on a physical Unitree G1 humanoid robot.

AINVIDIA reports that fine-tuned Nemotron models reached gold-medal level at both IOI 2026, scoring 535.4 out of 600, and IMO 2026, scoring 30 out of 42. The IOI run was a live, unofficial, unsupervised benchmark, while IMO proofs were graded by official IMO graders. The post also releases checkpoints, datasets, a new 200-problem benchmark, and inference pipelines on Hugging Face and NeMo-Skills.

Why it matters: The post traces how SFT, RL, and a generate-verify-refine loop turned Nemotron into gold-level specialists for IOI and IMO, with the training and inference details shared.

AIOrange AI released "podcast-to-article," a free MIT-licensed open-source skill that reorganizes podcast transcripts by topic rather than speaker timeline. The author says it keeps every number and original quote while cutting off-topic tangents, and that early readership and engagement were good.

AIThe final post in O'Reilly Radar's four-part post-training series walks readers through implementing the classic ChatGPT pipeline on Qwen2.5-1.5B, covering SFT, reward model training, and PPO. The walkthrough uses torchtune for SFT and verl, a Ray-based RL framework from ByteDance's team, for reinforcement learning. The author says the goal is hands-on understanding rather than reproducing InstructGPT, which took a large team and thousands of GPU-hours.

AIAi2 has published its Bolmo byte-level language model research in Nature and released new checkpoints on Hugging Face. The byteifying process converts an existing subword model into a byte-level one with a relatively short additional training run, and the paper reports that it also works for Qwen 3 8B and Llama 3 8B, producing Bwen 8B and Blama 8B. Ai2 also released Stage 1 checkpoints for researchers extending the architecture.

AIThe xAI Cookbook now has ten runnable Grok API examples across three tracks: real-time voice agents, multimodal generation, and live X data analysis. The author says four voice examples show the same Realtime Voice API across WebSocket, WebRTC, Twilio phone, and mobile transports. The four multimodal examples chain understanding, image generation or editing, video, and TTS, with Grok making creative decisions and Imagine models executing them.

AIPaige Bailey, a Google and Gemini figure, invited holders of high-quality, specialized data to license or sell it to improve Google products through a new portal, contentpilot.google.com. The post frames data as the most important asset and welcomes such partnerships, but gives no terms, pricing, or eligibility details.

AIFrançois Chollet asks whether the jagged frontier of AI capability is mainly math and code, which can be pushed far with RLVR. He questions whether steady gains in non-verifiable areas come from higher generalization driven by RLVR or only from continued injection of new human data.

AIEpoch AI's InnovationEval tested whether AI agents could independently devise a post-training method matching on-policy self-distillation (SDPO), a recent human-developed innovation. GPT-5.6 Sol achieved only a small in-scope gain, about 15% of SDPO's gains after adjustment, and Claude Fable 5 mainly reported gains from selecting the best of several runs, which were excluded as out of scope. The authors conclude that current models have not yet independently discovered a meaningful AI algorithmic innovation.

Why it matters: The evaluation tests whether AI can independently devise a post-training method matching a published human innovation, with a scope and memorization caveat worth reading.

AIEpoch AI reports that the share of US adults reporting at least one cyber incident in the past 12 months was 45% in September, essentially unchanged from 46% in June. The poll found no detectable change among frequent AI users, who moved from 53% to 51%. Epoch notes that its polling measures ordinary Americans' experiences, separate from its documented rise in serious vulnerability disclosures and frontier-model offensive capabilities.

AIPyTorch's blog describes a Triton-based implementation of Table Batched Embedding (TBE) forward and backward kernels for recommendation-system embedding lookups, which the post says outperforms legacy CUDA kernels on these workloads. On B200, an updated CUDA bounds-check step reaches up to 1.24x speedup on that component, and an optional forward-side preprocessing path cuts combined latency from 79.537 ms to 66.183 ms (−16.8%) on a large configuration.

AIFireworks' Head of AI Developer Education fine-tuned a 9B Jev-style classifier on Fireworks using plain SFT, public datasets, and two simple tricks. The full recipe is open source so developers can train their own decision classifiers without a frontier lab budget.

AIGoogle Earth AI combines environmental signals and other data sources with AlphaEarth Foundations, a Population Dynamics Foundation Model (PDFM), and a prototype Geospatial Reasoning agent. Researchers ask questions such as where a disease is likely to spread next, and the system automatically gathers relevant models and datasets to build a prediction model. By combining satellite views with population patterns, the tool aims to reveal hidden risk factors and identify issues earlier.

AIGoogle Earth AI, according to new research, can help communities respond to public health crises more quickly and proactively. The post says it combines behavioral trends, geospatial AI models, and other insights beyond simple statistics to help public health teams understand complex issues and bridge reporting gaps. The aim is to shift emergency response from reactive management toward proactive prevention.

AIZyphra trained its ZAYA1-8B reasoning model from scratch on a full-stack AMD platform, according to AMD's post. VP of AI Engineering Quentin Anthony credits access to open software libraries and direct collaboration with AMD for enabling bigger model training and efficient compute use.

AIMicrosoft has been named a Leader in the 2026 Gartner Magic Quadrant for Global Industrial AIoT Platforms. The company says its Azure platform, including Azure IoT, Azure Arc, Microsoft Fabric, and Microsoft Foundry, connects cloud and edge operations to apply AI-powered reasoning and close the loop between insight and action.

AINVIDIA's technical blog describes bitwise determinism for large-scale pretraining with Megatron Core, which makes training runs easier to debug, validate, and resume reproducibly. The source says these benefits matter most for models with trillions of parameters trained across thousands of GPUs, where multiple parallelism dimensions, low-precision computation, and distributed checkpointing complicate failure reproduction and fix validation.

AIParsewave audited all 600 public tasks in Zapier's AutomationBench and human review confirmed 206 real verifier bugs, all of which were fixed in AutomationBench Verified. Replaying 1,235 Kimi K3 runs on the old and fixed verifiers changed 27.9% of grades, with pass rate rising from 18.8% to 43.8% where verifiers were too strict and falling from 60.2% to 49.7% where they were too lenient.

AIChina's booming artificial intelligence sector has kept local researchers at home, but it has not yet attracted overseas scientists despite a government push to recruit them.

AIAi2 applied scientist Shashank Gupta says he will attend COLM 2026 from Tuesday through Friday and invites people to talk with the Ai2 Agents post-training team. Topics include post-training for coding and long-horizon agents, such as agentic RL, OPD, data and infrastructure, and multi-agent training, plus opportunities at Ai2. He lists an Ai2 booth session Tuesday 1:30–3pm and an Ai2 mixer Tuesday 6–9pm.

AIMicrosoft Research podcast host Chad Atalla interviews Jennifer Neville, a partner research manager at Microsoft, about her path into AI and her work on how evaluation exposes surprising failures in models tested beyond traditional benchmarks. The conversation also covers practical guidance for working with current AI systems and why examining underlying data matters when results defy expectations.

AIGoogle Research shares results from five partner-driven case studies showing how Google Earth AI's Population Dynamics Foundation Model addresses public health data gaps. The post frames the work as empowering global health research, with details available in the linked blog.

AIGoogle Research reports that Population Dynamics Foundation Model (PDFM) embeddings, built from search trends, mobility, built environment, and weather signals, were tested by partners across five public health tasks. The embeddings improved results in cross-border MMR vaccination coverage, dengue forecasting, postpartum depression screening, and cholera outbreak prediction, and matched census inputs for cardiovascular mortality nowcasting.

AISemiAnalysis rated FarmGPU as ClusterMAX Underperform after its Slurm layer failed to advertise GPU resources and Kubernetes exposed no RDMA devices for scale-out networking. The post credits FarmGPU's Grafana monitoring, provisioning notes, and trustworthy technical team, while noting the team may be stretched thin across small clusters.

AICisco SVP Rakesh Chopra discusses the "Scale-Across" approach to networking AI training workloads spread across multiple data centers. He describes how Silicon One architecture and Intelligent Collective Networking aim to manage synchronized GPU traffic over long-distance fiber links. The interview covers power efficiency and hardware-accelerated MACsec and IPsec security.