NewsroomProducts & Tooling
Products & Tooling
Apps, platforms, developer tools, open-source releases, and pricing changes.
Apps, platforms, developer tools, open-source releases, and pricing changes.
NVIDIA has published NemotronLabs VoiceChat, an 11-billion-parameter end-to-end speech model the company describes as the first open full-duplex voice model that can call tools mid-conversation,…
Replit dropped prices across its Cloud platform on August 1, cutting Autoscale compute costs by more than 80% and database storage costs by more than 75%, with the new rates applied automatically to…
The Allen Institute for AI (Ai2) has launched the OlmoEarth Platform, open infrastructure designed to take its OlmoEarth Earth-observation models from research fine-tuning to running inference across…
Liquid AI has released two new open-weight encoder models, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, aimed at classification and routing workloads that need to run continuously and cheaply on CPUs…
A Google creative technologist named Raph has built and open-sourced Glanceboard, a family scheduling tool that turns calendar and weather data into a personalized illustration shown on an e-ink…
OpenAI disclosed Friday that its models now reach more than one billion active users and more than two million businesses, one of the company's most specific public statements yet about the scale of…
OpenAI slashed API pricing for two of its three GPT-5.6 models on July 30, cutting the cost of GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20%, according to the company's official API changelog. What's…
OpenAI announced ChatGPT for Academic Researchers on July 29, a program offering free access to its frontier models and research tools to academics, starting with 10,000 researchers this summer and…
Google has introduced Gemini Spark, an autonomous personal AI agent built on Gemini 3.6 Flash, to users in India, expanding a feature it first previewed at Google I/O 2026 into general availability…
Cohere has introduced North Automations, a new orchestration layer for its North enterprise AI platform that lets employees describe a multi-step business process in plain language and have several…
Moonshot AI has published the full weights of Kimi K3 on Hugging Face, delivering on the timeline it set 11 days ago when it first unveiled the 2.8-trillion-parameter model. The model card, now live…
The Swiss AI Initiative — a joint effort of ETH Zurich, EPFL, and the Swiss National Supercomputing Centre (CSCS) — released Apertus 1.5, an update to its fully open language model family that adds…
xAI has shipped Grok for Excel, a free Microsoft 365 add-in that puts its Grok assistant directly inside spreadsheets. "Today we're bringing Grok into Microsoft Excel," the company said in its…
NVIDIA has open-sourced a Medical Physics Simulation framework, a GPU-accelerated capability within NVIDIA Isaac for Healthcare that lets medical robotics developers simulate anatomy-device…
OpenAI has added organization- and project-level spend limits to the API platform, letting developers cap monthly API costs or enforce hard limits that fail requests once a limit is reached. What's…
Poolside released Laguna S 2.1 on July 21, 2026, an open-weight foundation model built for agentic coding and long-horizon software engineering tasks, which the company is positioning as the West's…
Meituan released LongCat-2.0, a 1.6-trillion-parameter mixture-of-experts language model, on Hugging Face under the MIT license — the latest and largest entry in its LongCat family, and one it says…
Moonshot AI unveiled Kimi K3 on July 16, 2026, a 2.8-trillion-parameter model the company calls the world's first open 3-trillion-class model. The API is live today on Kimi.com, Kimi Work, Kimi Code,…
Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, released Inkling on July 15, 2026 — its first open-weight model and a direct challenge to the "one-size-fits-all"…
Google is renaming NotebookLM to Gemini Notebook, folding its source-grounded research tool more tightly into the Gemini ecosystem while giving it a significant new capability: a secure cloud…
A German research consortium has released Soofi-S-30B-A3B, an open-weights foundation model built as a sovereign European alternative to US and Chinese AI systems, with training run entirely on…
ElevenLabs formally launched its Canadian business on July 7, 2026, appointing a dedicated General Manager, committing to double its Canadian headcount this year, and opening its first Canadian…
Anthropic launched Claude for Teachers on July 14, 2026, giving verified K-12 educators in the US free access to premium Claude capabilities built around lesson planning, differentiation, and…
NVIDIA published Nemotron-3-Embed-8B-BF16 on Hugging Face on July 14, 2026, an open-weights text embedding model built for retrieval-augmented generation (RAG) and semantic search across 34…
Together AI introduced Provisioned Throughput on July 8, 2026, a reserved-capacity purchasing option aimed at teams running open-weight models in production who need guaranteed performance rather…
OpenAI's head of safety systems, Johannes Heidecke, is leaving the company, following a reorganization that merges OpenAI's safety and research teams under a single leader. What's new Per Wired,…
Meta shut down a feature that let people @-mention any public Instagram account inside Meta AI and generate images referencing that account's photos, three days after it shipped, following objections…
Cohere has open-sourced Cohere Transcribe Arabic, a 2B-parameter automatic speech recognition model purpose-built for Arabic dialects, code-switching, and business/developer speech, releasing the…
Salesforce has made its Agentforce Commerce suite of AI shopping agents generally available, with Shopper Agent, Buyer Agent, and Merchant Agent now live and natively integrated into ChatGPT, ahead…
Fidji Simo, OpenAI's CEO of Applications and the company's second-highest-ranking executive, announced on July 9, 2026 that she is stepping back from her full-time role to focus on recovering from a…
OpenAI is shutting down Atlas, the standalone AI-powered browser it launched less than a year ago, and moving its agentic browsing features into the core ChatGPT desktop app and a Chrome extension.…
OpenAI has launched ChatGPT Work, a new agent inside ChatGPT designed to take on a goal, work independently across a user's connected apps and files for hours if needed, and hand back finished output…
Anthropic has launched a public beta of Claude Code and Claude Cowork for Claude for Government Desktop, a FedRAMP High authorized offering built for U.S. federal, state, and local government…
Hugging Face has released version 0.6.0 of LeRobot, its open-source robot-learning library, headlined by a new class of "world model" policies that predict what will happen next before a robot…
Z.ai, the Beijing-based lab formerly known as Zhipu AI, has launched ZCode, a free desktop coding environment built specifically around its flagship GLM-5.2 model. The release, which rolled out the…
Meta CEO Mark Zuckerberg told employees at an internal town hall on July 2, 2026 that the company's AI agent development has not moved as fast as leadership expected, an unusually candid admission…
Mistral has released Leanstral 1.5, a free Apache-2.0 licensed model built specifically for formal verification and proof engineering in Lean 4, and it has effectively maxed out one of the field's…
Meta has quietly rate-limited Conversation Focus, an on-device AI feature on its smart glasses that amplifies a speaker's voice in noisy settings, capping free use at three hours a month and tying…
Chinese delivery-app giant Meituan open-sourced LongCat-2.0 on June 30 — a 1.6-trillion-parameter coding model it says is the first of its scale trained and served end-to-end on domestic Chinese…
Meta has quietly launched Pocket, an experimental app that lets users generate and play small interactive games — which it calls "gizmos" — from a single text prompt, without any official company…
Microsoft has launched Microsoft Frontier Company, a new operating unit that embeds 6,000 engineers and industry specialists directly inside enterprise customers to design, deploy, and continuously…
DeepSeek has released DSpark, an add-on speculative-decoding module for its DeepSeek-V4 models, and published it on Hugging Face under an MIT license alongside a companion training codebase called…
Microsoft's AI For Good Lab partnered with the Theodore Roosevelt Presidential Library to build a set of AI tools that let visitors search Roosevelt's writings in plain language and converse with a…
Suno has introduced Spark, an incubator program that gives independent musicians grants and marketing support to make and promote new music on the platform — while binding participants to a clause…
Cursor released a native iOS app in public beta, letting developers launch and steer AI coding agents from their phone instead of only from a desktop editor. What's new Cursor for iOS is available…
Hugging Face and Cerebras have released an open, cascaded speech-to-speech voice AI demo built on Google DeepMind's Gemma 4 31B, pairing Cerebras's inference speed with a fully open, swappable model…
NVIDIA released Nemotron 3 Nano Omni on April 28, 2026, an open multimodal model designed for AI agent deployments that unifies video, audio, image, and text processing in a single architecture. The…
Amazon Web Services launched a new Forward Deployed Engineering (FDE) organization on June 30, 2026, committing $1 billion to place AWS engineers directly inside enterprise customer organizations to…
Anthropic launched Claude Science on June 30, 2026, a dedicated AI workbench for scientific research that integrates more than 60 scientific databases into a single environment, enabling researchers…
Krea released the weights for its Krea 2 image generation models on June 23, 2026, making K2 Raw and K2 Turbo available for download under a permissive license. The release marks Krea's entry into…
NVIDIA launched Ising, a family of open-source AI models designed to serve as the control plane for quantum processors — the first such models publicly released. Ising addresses two foundational…
DeepSeek published DSpark on June 27, 2026 — a speculative decoding framework applied as an add-on module to its flagship DeepSeek-V4-Flash and V4-Pro models, boosting per-user generation speed by…
Liquid AI on June 25, 2026 released LFM2.5-230M, its smallest model yet — a 230-million-parameter non-transformer model that runs at 42 tokens per second on a Raspberry Pi 5 and outperforms models…
NVIDIA announced the Nemotron 3 family of open models on June 4, 2026 — a three-tier lineup built on a hybrid mixture-of-experts (MoE) architecture that uses NVIDIA's 4-bit NVFP4 training format on…
Google unveiled Gemini Spark at its I/O developer conference on May 19, 2026 — a persistent, cloud-based AI agent designed to handle multi-step tasks across Gmail, Google Workspace, and third-party…
Anthropic on June 26, 2026 raised rate limits across the Claude API and collapsed its usage tier structure to three buckets — Start, Build, and Scale — eliminating the previous multi-tier ladder that…
Four of Google's most prominent AI researchers departed in a single week for rival labs, with three heading to Anthropic and one to OpenAI. The exits — including a Nobel laureate and one of the…
Nex AGI has released Nex-N2-Pro, a 397-billion-parameter open-source Mixture-of-Experts model under an Apache 2.0 license, alongside a smaller Nex-N2-mini variant. Built on the Qwen3.5-397B-A17B…
Google has open-sourced DiffusionGemma, an experimental 26-billion-parameter Mixture of Experts text-generation model that ditches token-by-token autoregressive decoding in favor of parallel block…
Google announced on June 15, 2026 that Veo 2.0 and Veo 3.0 — its current video generation models available via the Gemini API — will shut down on June 30, 2026. The same changelog entry confirmed…
Cohere released Command A+, its most capable model to date, on May 20, 2026 under an Apache 2.0 open-source license. The 218-billion-parameter Mixture-of-Experts model is designed to run on minimal…
OpenAI has announced DevDay 2026, its annual developer conference, scheduled for September 29 at Fort Mason in San Francisco. Applications are open now through July 10, 2026, with accepted applicants…
Anthropic today launched Claude Tag in beta, a new product that adds Claude as a persistent member of Slack workspaces. Unlike existing point-in-time AI integrations, Claude Tag gives every member of…
Google announced on June 15, 2026 that six production video and image generation models in the Gemini API are being retired, with the first group -- three Veo models -- shutting down on June 30,…
NVIDIA has published a comprehensive framework for enterprise AI agent development under the NVIDIA Agent Toolkit banner, consolidating its Nemotron open models, NemoClaw agentic blueprints, and…
Zhipu AI's Z.ai released GLM-5.2 on June 13, 2026, a 753-billion-parameter mixture-of-experts model that outperforms GPT-5.5 on multiple long-horizon coding benchmarks while costing roughly one-sixth…
NVIDIA unveiled a comprehensive suite of products on June 22 designed to bring autonomous, 24/7 AI agents to telecommunications network operations, debuting the platform at DTW Ignite 2026 in…
Sakana AI launched Fugu on June 22, 2026, a new kind of AI model designed to orchestrate a dynamic pool of large language models through a single OpenAI-compatible endpoint. Unlike a traditional…
Google has announced end-of-life dates for several of its Gemini API video and image generation models, with Veo 2.0 and Veo 3.0 shutting down on June 30, 2026 — eight days away — and Imagen 4 models…
NVIDIA released Nemotron 3 Ultra on June 4, 2026 — a 550 billion total parameter, 55 billion active parameter open-weights model built for long-horizon agent workloads. The model uses a hybrid…
Andrej Karpathy, one of the founding members of OpenAI and one of the most recognized AI researchers in the world, announced on May 19, 2026 that he has joined Anthropic. He will work on the…
DeepSeek on April 24, 2026 released DeepSeek-V4-Pro and DeepSeek-V4-Flash simultaneously — two open-source Mixture-of-Experts models that bring 1-million-token context windows, a new sparse attention…
Google announced the Google Home Speaker on June 17, 2026 — its first audio device in six years and the first purpose-built to run Gemini as its voice assistant. Priced at $99.99, it is available for…
Google opened Dataland on June 20, 2026 in Los Angeles — a 25,000-square-foot interactive museum billed as the world's first institution dedicated to AI-generated art. Built in The Grand LA, Frank…
Z.AI has released GLM-5.2, its latest open-weight flagship model for long-horizon tasks. The 753B-parameter model is available under an MIT license with no regional restrictions and ships with a…
OpenAI on June 14, 2026 launched its first formal partner program — the OpenAI Partner Network — backed by a $150 million investment designed to build a global ecosystem of certified AI…
Amazon Web Services launched the AWS FinOps Agent in public preview on June 9, 2026, an AI-powered agent designed to help engineering and finance teams investigate cloud cost anomalies, answer cost…
The City of Rio de Janeiro has released Rio 3.5 Open 397B, a 403-billion-parameter open-weights AI model built by IplanRIO, the municipal technology company that manages the city's digital…
Shanghai-based AI lab MiniMax released MiniMax-M3 on June 1, 2026 — a 427-billion parameter open-weight model with a 1-million-token context window, native multimodal understanding, and a new sparse…
DeepSeek has released DeepSeek-V4-Pro and DeepSeek-V4-Flash on Hugging Face, its most capable open-weight models to date. The V4-Pro variant is a 1.6 trillion total parameter mixture-of-experts model…
NVIDIA unveiled Cosmos 3 on May 31, 2026 at GTC Taipei — an open world foundation model for physical AI capable of natively understanding and generating text, images, video, ambient sound, and…
NVIDIA on June 4, 2026, launched Nemotron 3 Ultra, an open-weights 550-billion-parameter reasoning model designed for enterprise AI agents. The model delivers claimed "5x higher throughput compared…
Mastercard launched Agent Pay for Machines on June 10, 2026 — an open payments protocol designed to let AI agents and autonomous systems execute high-frequency, low-value transactions…
Moonshot AI has released Kimi K2.7 Code as an open-weights model on Hugging Face, extending the Kimi K2 lineage with a coding-specialist architecture that packs 1 trillion total parameters into a…
Cohere on June 9, 2026 released North Mini Code — the company's first agentic coding model and the opening release of its next generation of models. The model uses a mixture-of-experts architecture…
Amazon Web Services published Agent-EvalKit on June 11, 2026 — an open-source toolkit released under the Apache 2.0 license that gives developers a structured pipeline for evaluating AI agents by…
Anthropic on June 11, 2026, launched Claude Corps, a national fellowship program placing 1,000 early-career workers at nonprofits across the United States over the next year, backed by an initial…
NVIDIA detailed the production architecture of its Halos Operating System at GTC Taipei on June 10, 2026, announcing four simultaneous commercial robotaxi deployments across Munich, Taiwan, Southeast…
Google DeepMind has released DiffusionGemma, an experimental open-weights language model that generates text using a diffusion process rather than token-by-token prediction, achieving speeds up to…
At WWDC 2026 on June 8, Apple introduced the third generation of Apple Foundation Models (AFM 3), a family of five purpose-built AI models co-developed with Google that spans on-device inference and…
Google DeepMind kicked off its first dedicated robotics accelerator on June 9, 2026, gathering 15 selected European startups in London to begin a three-month program built around access to the…
Google shut down all four Gemini 2.0 Flash model variants on June 1, 2026, completing the retirement of the 2.0 Flash generation and clearing a path for Gemini 3.5 Flash as the standard…
OpenAI updated its container session billing on June 2, 2026, moving from a flat 20-minute session rate to per-minute billing with a 5-minute minimum. The change reduces costs for developers whose…
OpenAI changed how container sessions are billed on the API, swapping the flat 20-minute session charge for per-minute billing with a 5-minute minimum. The shift trims cost for short-running…
Gemini 3.5 Flash is priced at $1.50 per million input tokens and $9 per million output tokens - 3x the $0.50 / $3 of Gemini 3 Flash, but still roughly 40% cheaper than Gemini 3.1 Pro. Google has…
GPT-5.5 became available to API developers on April 24, 2026 at $5 per million input tokens and $30 per million output tokens with a 1M-token context window. GPT-5.5 Pro is priced at $30 per million…