Releases

Perplexity launched Sandbox API for isolated code execution by AI agents.

Perplexity upgraded Search API with improved extraction and dynamic evaluation benchmarks.

Perplexity launched Agent API as a managed runtime for agentic workflows.

Meta published an infrastructure update describing four MTIA chips developed across two years for large-scale AI inference.

NVIDIA released the open 120-billion-parameter Nemotron 3 Super model for complex agentic AI systems.

CoreWeave introduced more flexible cloud-capacity plans designed to match customer workload duration and deployment needs.

Hugging Face launched mutable, Xet-backed S3-like Storage Buckets for checkpoints, logs, traces, and intermediate AI data.

SpaceXAI released Grok 4.20 and a multi-agent variant through its platform.

Meta and World Resources Institute released a DINOv3-based global forest-canopy model with improved accuracy and detail.

Harvey evolved Workflow Builder into Agent Builder for reusable custom legal agents.

Harvey reported hundreds of thousands of agentic Deep Analysis reports and expanded its use in custom agent workflows.

Databricks redesigned AI Search with decoupled storage and compute to support billions of vectors at lower cost.

Scale Labs introduced repository-grounded rubrics for evaluating and reranking candidate software patches without executing tests.

Scale Labs introduced VeRO, a reproducible harness using versioned agent snapshots, controlled budgets, structured traces, and reference procedures.

Scale expanded SEAL into Scale Labs, a broader research organization focused on evaluation, agents, multimodal systems, post-training, deployments, and oversight.

OpenAI released GPT-5.4 and GPT-5.4 Pro for professional work, coding, computer use, and long-context tasks.

Harvey announced an embedded legal-intelligence agent for Microsoft 365 Copilot.

Modal launched Directory Snapshots for capturing and restoring sandbox filesystem state.

Modal added integrated changelog visibility and workflow improvements to its CLI.

CoreWeave announced an agreement to provide AI inference infrastructure for Perplexity.

Cohere joined Aston Martin Aramco as official generative-AI partner to support engineering and business innovation.

Scale launched SWE Atlas, initially releasing Codebase QnA to evaluate how coding agents investigate and reason about real software systems.

Microsoft Research releases Phi-4 Reasoning Vision, a 15-billion-parameter open-weight multimodal reasoning model.

Perplexity expanded Computer as the unifying interface for search, research, creation and actions.

AWS made centralized policy controls for agent-tool interactions generally available.

Google launches Gemini 3.1 Flash-Lite for low-cost, high-throughput multimodal workloads.

Moonshot published its distributed model-checkpointing engine.

Moonshot published optimized kernels for Kimi Delta Attention.

Moonshot released its expert-parallel communication library for large MoE training and inference.

Moonshot published Mixture-of-Block Attention code and research for efficient long-context modeling.

Baidu launched a general-purpose agent for multi-step productivity workflows.

Scale launched high-fidelity simulated environments for training and testing agents on tool use, computer use, coding, and professional workflows.

Amazon and OpenAI announced joint development of a stateful environment for long-running agents on Amazon Bedrock.

Amazon and OpenAI announced that next-generation Trainium4 is expected to begin delivery in 2027.

Amazon announced an initial $15 billion investment in OpenAI followed by up to $35 billion more under preset conditions.

OpenAI and Amazon announced a strategic partnership spanning investment, AWS compute, and distribution of OpenAI models and products.

OpenAI announced $110 billion of new investment from Amazon, NVIDIA, and SoftBank at a $730 billion pre-money valuation.

Midjourney expanded web personalization profiles and moodboard workflows.

Cursor released Composer 1.5 with stronger agentic coding performance and positioned proprietary models as central to its product strategy.

CoreWeave reported fiscal 2025 revenue of $5.131 billion, up 168% year over year, with $66.8 billion of revenue backlog.

Samsung partnered with Perplexity to integrate its APIs and AI search capabilities into Samsung devices.

Google launches Gemini 3.1 Flash Image for fast, controllable image generation and editing.

Perplexity launched Computer, an agentic workspace that coordinates models, web search, tools, connected apps and code to complete knowledge-work tasks.

NVIDIA reported Q4 fiscal 2026 record revenue of $68.1 billion, up 73% year-over-year and 20% sequentially. Data Center revenue was $62.3 billion. Full-year FY2026 revenue reached $215.9 billion (+65% YoY), with full-year net income of $120.1 billion (GAAP).

Cursor Cloud Agents gained computer-use capabilities for interacting with browsers and graphical applications.

Anthropic published version 3.0 of its Responsible Scaling Policy, updating its voluntary framework for managing catastrophic risks from advanced AI.

Google launches Gemini 3.1 Pro with advances in reasoning, coding and long-running agentic tasks.

Mistral AI agreed to acquire French serverless cloud platform Koyeb to strengthen its AI infrastructure and deployment stack; financial terms were not disclosed.

Anthropic released Claude Sonnet 4.6 with upgrades across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.

Cursor introduced long-running agents designed to execute extended software-engineering tasks in the cloud.