AWS Machine Learning Blog3434OpenFold3-preview can be deployed on Amazon SageMaker JumpStart, an open-source biomolecular structure prediction model developed by the OpenFold Consortium and AlQuraishi Lab at Columbia University, supporting 3D structure prediction for proteins, DNA, RNA, small molecule ligands, and their complexes.
AWS Machine Learning Blog4545Amazon Bedrock AgentCore runtime V2 reduces session startup time to about 2 seconds by preparing the environment once and using a snapshot for recovery, while also reclaiming unused memory. V2 supports building hooks to load large files or CLI, suitable for scenarios requiring resource preparation in advance. Session startup time can be measured via CloudWatch metrics, prioritizing optimization of platform overhead or code execution time.
AWS Machine Learning BlogDevelopers using coding agents like Claude Code and OpenAI Codex CLI need to be billed per task rather than per single call, so model routing should be based on the cost per task. NVIDIA NeMo Switchyard is an open-source library that helps AI agents select models to process each request, supporting Stage and Random routing strategies.
Codex releases00GitHub Copilot released version 0.171.0-alpha.5, providing code review, vulnerability detection, and secret scanning features, supporting enterprise-level code security and AI features.
Google ResearchFeatured8888Google has opened the /teamwork multi-agent workflow in Antigravity to external researchers and developers, which can be used in conjunction with Gemini 4 Argon, Carbon, and Hydrogen; /boost and Deep Review are also available.
Why it matters: Google has disclosed /teamwork's specific results in multi-agent research and engineering tasks and made it available for external researchers to use.
LangChain Deep Agents releases1212langchain-modal 0.0.8 has been released, fixing the `ModalSandbox` file I/O issue and updated to use the `Sandbox.filesystem` API
AWS Machine Learning Blog3838Amazon Bedrock tracks every inference call to an IAM principal with built-in zero instrumentation cost attribution, passing identity information and organizational tags to AWS Cost and Usage Reports (CUR 2.0) and AWS Cost Explorer while keeping data within AWS security boundaries.
AWS Machine Learning BlogAssemblyAI builds an elastic architecture on AWS by hybridly using EC2 On-Demand and Spot instances, and leveraging Amazon ECS, SQS, and EventBridge, to increase GPU instance utilization by 3.5 times, achieve cost reduction within the European AWS region, and meet service level agreements (SLA). Its inference platform is deployed on three GPU instance families: g4dn, g5, and g6.
Ai2 BlogSame story · 2 outlets4040XYBench released, including 8K+ problem tests to evaluate how AI assistants handle requests based on misunderstandings; Claude Sonnet 4.6, GPT-5, Gemini 3.1 Pro, Qwen3.5-9B, and DeepSeek-V4 Pro often focus on the original incorrect method rather than the user's actual goal in the test.
How the accounts differ:Ai2 Blog states the benchmark includes 8K+ problems, while Hugging Face Blog mentions 8150 problems.
Meta EngineeringMeta has natively ported the Android Virtual Device framework Cuttlefish to macOS, by supporting QEMU and vhost-user, solving the technical challenges of running virtual devices on Apple Silicon, improving development efficiency, and unifying testing environments. The new version supports macOS's display, signing, and build processes, and expands features such as Bluetooth, Wi-Fi, and 3D graphics, allowing Mac developers to use high-fidelity virtual devices the same as CI.
Hugging Face BlogSame story · 2 outlets4040Hugging Face releases XYBench benchmark, containing 8150 real-world problems, to evaluate AI assistants' ability to identify user misunderstandings and guide them to correct solutions. Claude Sonnet 4.6, GPT-5, Gemini 3.1 Pro, Qwen3.5-9B, and DeepSeek-V4 Pro tend to focus on user errors (Y) rather than actual goals (X), with only 35-41% of responses centered on the actual goals.
How the accounts differ:Ai2 Blog states the benchmark includes 8K+ problems, while Hugging Face Blog mentions 8150 problems.
Databricks BlogDatabricks has released Genie Migrate, now supporting migration to Lakehouse from Microsoft SQL Server, Azure Synapse, Teradata, Oracle, Snowflake, Amazon Redshift, and Google BigQuery, automatically converting SQL code to ANSI SQL.
LangChain Deep Agents releases2020LangChain Deep Agents 0.7.38 was released to fix the issue where `ls`, `grep`, and `glob` would incorrectly return results when soft link resolution exceeds the approved search scope.
Ant Group inclusionAI · Hugging Face new modelsinclusionAI releases the Duo 1.0 Realization weights for FP8-quantized four-step distillation, used for its causal base model, supporting full-duplex real-time audio-video interaction, and needs to be overlaid with the causal base model weights.
Codex releases00GitHub Copilot 0.171.0-alpha.2 has been released, the AI coding agent is used for tasks, projects, and automation work, supporting code review, collaboration, and cloud-based development environments.
Ant Group inclusionAI · Hugging Face new modelsSingVL-4B GGUF is the GGUF version of the 4B parameter multimodal decision-making model Sing VL S1 based on Qwen3.5, which can answer named entity questions with text or JSON and up to four images, and uses a full forward pass scoring option, scoring 69.76 in the Text evaluation and 73.94 in the Vision evaluation. The model uses the llama.cpp adapter, with a default context size of 4096 tokens, supporting text and image input.
Databricks Blog2222Lakebase architecture serves both Postgres and analytics with a single persistent data copy, without the need for CDC pipelines or a second database. This architecture stores data hierarchically in object storage and cache, not traditional replication. It supports both point-in-time and scan access paths, optimized respectively using Postgres heap pages and Parquet format, with no additional synchronization tasks.
vLLM BlogDeepSeek-V4.1-Flash achieves a service capability of 1.7M tokens/s per GPU in a single NVIDIA Vera Rubin NVL72 rack, reaching up to 387 tokens/s per user, an increase of 58 times compared to GB300 NVL72.
AWS Machine Learning BlogNatera developed the SWARM intelligent agent system to address knowledge gaps caused by laboratory equipment failures, achieving 130x response speedup and 96.2% automation support for handling work orders through Amazon Bedrock AgentCore.
AWS Machine Learning BlogSGLang-Neuron made available via AWS Deep Learning Container, supporting SGLang inference and reinforcement learning deployment on AWS Trainium, using the same interface as SGLang, running on native PyTorch, supporting PD disaggregation and 10240 token context length.
vLLM BlogFeatured6060vLLM provides support for Inkling 2 Small on its first day of release. The model is a multimodal MoE with text and visual inputs, featuring a total of 594B parameters, 17B activation parameters, a maximum context length of 1M tokens, and comes with a built-in DSpark beam decoder.
Why it matters: The original text provides details of vLLM's Day-0 support for Inkling 2 Small and multiple kernel optimizations, allowing readers to understand the deployment path and throughput performance of the model.
AWS Machine Learning BlogAmazon Bedrock's request-level selection still changes the number of tokens processed by the service and the applicable tax rate. Cache duration, service tier, inference effort, and maxTokens can affect costs without changing the model or prompt. Each selection depends on the model and workload, and its effectiveness can be verified through response metadata, logs, and billing data.
Google Developers BlogGemini 4 Argon translates Go code panic paths into Z3 constraints, paired with Z3 to check trigger conditions, and paired with go test to replay trigger bytes, locating 3 embedded parser bugs and discovering 3 skill-lint potential bugs within 1 minute. This process is more efficient than go test -fuzz, capable of finding boundary check errors missed by fuzz testing.
AWS Machine Learning BlogAmazon Bedrock automatically records caller identity, users can assign tags through IAM hosts, activate cost allocation tags, and use CUR 2.0 or AWS Cost Explorer to report data, combined with AWS Budgets to set budget actions for spending limits. No application code needs to be written, when the threshold is reached, budget actions will automatically apply Deny IAM policies. This solution is suitable for accounts already using IAM roles, supporting both static primary tags and dynamic session tags for identity mapping.
Databricks BlogEclipse Aerospace partnered with Databricks to make nearly 20 years of proprietary flight data (totaling 3TB) more accessible for service centers and engineers via a custom data source using PySpark, without requiring additional infrastructure.
Codex releases00GitHub Copilot has released version 0.170.0-alpha.4, providing code review, security vulnerability detection, and key scanning features.
AWS Machine Learning Blog3838Developers can quickly access base models through Amazon Bedrock application inference profiles and LiteLLM, achieving single API access, user-level cost control, and model performance observability.
Google AI BlogFeatured8282Google Search has launched Gemini 4 Hydrogen, the largest model upgrade to date, now available to all global users. The company claims the new model can understand multi-part questions, spelling errors, and highly specific requests, and leverages stronger instruction following and tool calling capabilities to respond to each question individually. Hydrogen's reasoning ability also enhances query distribution, allowing it to retrieve multiple related subtopics simultaneously and provide more images, structured responses, comparison tables, and custom tools.
Why it matters: Google's official statement on Gemini 4 Hydrogen explains changes in intent understanding, query distribution, and visual responses in search, which can be used to judge the direction of adjustments to the search experience.
OpenAI Alignment BlogOpenAI uses automated red team techniques to train GPT-6 Astra and Sol, significantly enhancing the model's fuzzy robustness against jailbreak attacks in high-risk areas such as networking, biology, and violent activities. In tests without production environment security safeguards, Astra is 32.5% safer than GPT-5.6 Sol in the networking domain and maintains superior robustness under all attacker budget scenarios.
Google DeepMindFeatured8787Google DeepMind releases Gemini 4 Hydrogen, targeting daily coding and intelligent agent workflows, with an Artificial Analysis intelligence index of 51 at the default high thinking level, higher than Argon's 50, and reaching 54 at the xhigh level
Why it matters: Google provides the division of labor and positioning of Hydrogen, Carbon, and Argon, as well as the actual differences in cost and speed between multi-model arrangements.
Apple Machine Learning Research3838Researchers derive the maximum expected reward gain of KL regularization alignment under a fixed KL divergence budget, for the first time describing optimal reward improvement using the Jeffreys divergence term rather than √KL; experiments show that commonly used methods like PPO and GRPO are far below the theoretical limit, and the best N-shot strategy approaches theoretical optimality, proposing algorithmic improvement directions for enhancing alignment performance.
AWS Machine Learning BlogAmazon Bedrock AgentCore runtime provides a secure environment for each session through hardware-isolated Firecracker micro-virtual machines (MicroVMs), the new version boosts startup speed by 15x, extends computing by 4x, improves efficiency by 68%, supports up to 8 vCPUs and 32 GB memory and 10 GB container images, session startup time is less than 2 seconds, suitable for complex workloads.
Databricks BlogH6.ai automates accounts payable processing on customers' existing ERP systems by reading invoices, matching purchase orders, allocating accounting subjects, and executing in accordance with company policies, combining AI-assisted understanding with deterministic policy execution. It adopts Lakebase Postgres to establish a dedicated database for each customer, storing purchase orders, suppliers, accounting subjects, historical transactions, and policies, etc., to support frequent regression testing.
PyTorch Blog5050Helion analyzes kernel IR and data flow through compile-time heuristic algorithms, selecting fast settings without slow automatic tuning, classifying kernels as matrix multiplication, reduction, or point-wise operations, and choosing block size, warps, and phases to achieve near-automatic tuning speed.
Databricks BlogBlock company has launched Buzz, an open-source collaborative space, allowing employees to interact with AI agents in a shared workspace. Through Databricks Unity Gateway, users can access cutting-edge models from OpenAI, Anthropic, Google, and open-source models (specific number not mentioned in the original text), experimenting with different AI agent framework combinations. Employees can create AI agents and choose models (such as the lightweight version of Goose) to achieve multi-person collaborative development, analysis, and other tasks.
AWS Machine Learning Blog5858Amazon Bedrock Managed Knowledge Base achieves fine-grained access control for enterprise AI agents by injecting user context into the AgentCore Gateway tool layer and verifying JWT, combined with the ACL-aware functionality of the knowledge base.
AWS Machine Learning Blog5555AWS introduces Agent-to-Agent (A2A) protocol, allowing multi-agent systems operating across organizational boundaries to discover each other, communicate, and exchange tasks through a standardized protocol without exposing internal details or relying on the same framework.
PyTorch BlogA team of two, with the help of an AI programming agent, completed the integration of SGLang and IBM Spyre in less than a week, building a plugin with about 1,500 lines of code. SGLang's inference throughput on Spyre is about 3.6 times higher than the reference PyTorch implementation (using torch-spyre) and performs comparably to vLLM on Granite 3.3 8B and Gemma 4 26B A4B workloads.
PyTorch BlogThe EUROPA Union successfully trained a MoE model with 951.9B parameters on the JUPITER supercomputer using PyTorch and TorchTitan, increasing single GPU throughput from 90 to approximately 1,300 tokens/s with 384 NVIDIA GH200 GPUs.
