j.AI
AI releases and models — what the vendors shipped, and what is running
What OpenAI, Anthropic, Google, xAI, Meta, Microsoft, Mistral and DeepSeek have announced, taken from their own feeds — and the models that are actually running, with context windows, prices and the date each one appeared. Nothing here is written by hand.
446models tracked
401announcements
17vendors
77new in 60 days
The field, tonight
Every model in the catalogue, placed by what it can hold and what it costs. Across: the context window, 4K to 2M. Up: the price of a million output tokens, three cents to six hundred dollars. Both scales are logarithmic or the picture is a smear. The larger, pulsing dots appeared in the last 30 days. Point at one.
5 routers quote no price at all — they charge whatever the model they pick charges — so they are the only things missing from the picture: Auto Router (Beta), OpenRouter: Fusion, Pareto Code Router, Body Builder (beta), Auto Router. They are counted everywhere else on this page. Leaving them out quietly would let the chart claim a completeness it does not have.
New models, last 60 days
Dated by when the model first appeared in the catalogue, not by when it was announced — an aggregator sees a launch hours to days late. Prices are per million tokens, in and out, as OpenRouter charges them; a vendor’s direct rate can differ.
| Appeared | Model | Vendor | Context | In | Out | Cutoff |
|---|---|---|---|---|---|---|
| 2026-09-18 | PrismML: Ternary Bonsai 2 27B | prism-ml | 262K | $0.07 | $0.50 | ā |
| 2026-09-18 | Z.ai: GLM 5.3 FlashX | z-ai | 1.0M | $0.37 | $1.25 | ā |
| 2026-09-17 | Pareto | unbiased | 262K | $2.50 | $7.50 | ā |
| 2026-09-14 | DeepSeek: DeepSeek Pro Latest | DeepSeek | 1.0M | $0.53 | $1.58 | ā |
| 2026-09-14 | DeepSeek: DeepSeek Flash Latest | DeepSeek | 1.0M | $0.13 | $0.52 | ā |
| 2026-09-12 | Inference.net: Schematron V2 Turbo | inference-net | 128K | $0.03 | $0.15 | ā |
| 2026-09-12 | Inference.net: Schematron V2 Small | inference-net | 128K | $0.05 | $0.23 | ā |
| 2026-09-11 | OpenAI: GPT Astra Latest | OpenAI | 1.1M | $10.00 | $50.00 | ā |
| 2026-09-11 | OpenAI: GPT Sol Latest | OpenAI | 1.1M | $2.00 | $10.00 | 2026-02-16 |
| 2026-09-11 | OpenAI: GPT Terra Latest | OpenAI | 1.1M | $2.00 | $12.00 | 2026-02-16 |
| 2026-09-11 | OpenAI: GPT Luna Latest | OpenAI | 1.1M | $0.20 | $1.20 | 2026-02-16 |
| 2026-09-11 | Sakana: Fugu Ultra v2 | sakana | 1.0M | $5.00 | $30.00 | 2026-08-28 |
| 2026-09-11 | Sakana: Fugu Max | sakana | 1.0M | $2.00 | $6.00 | ā |
| 2026-09-10 | inclusionAI: Ling 3.0 Flash VL | inclusionai | 131K | $0.06 | $0.18 | ā |
| 2026-09-10 | inclusionAI: Ling 3.0 Flash VL (free) | inclusionai | 262K | free | free | ā |
| 2026-09-10 | DeepSeek: DeepSeek V4.1 Flash | DeepSeek | 1.0M | $0.15 | $0.60 | ā |
| 2026-09-08 | Inception: Mercury 2.5 | inception | 260K | $0.04 | $0.15 | ā |
| 2026-09-08 | Nex AGI: Nex-N2.5-Mini (free) | nex-agi | 262K | free | free | ā |
| 2026-09-08 | Nex AGI: Nex-N2.5-Pro (free) | nex-agi | 262K | free | free | ā |
| 2026-09-04 | OpenAI: GPT-6 Astra | OpenAI | 1.1M | $10.00 | $50.00 | ā |
| 2026-09-04 | OpenAI: GPT-6 Astra (batch) | OpenAI | 1.1M | $5.00 | $25.00 | ā |
| 2026-09-04 | OpenAI: GPT-6 Astra Pro | OpenAI | 1.1M | $10.00 | $50.00 | ā |
| 2026-09-04 | OpenAI: GPT-6 Astra Pro (batch) | OpenAI | 1.1M | $5.00 | $25.00 | ā |
| 2026-09-04 | inclusionAI: Ling 3.0 Flash Sante (free) | inclusionai | 262K | free | free | ā |
| 2026-09-03 | Qwen: Qwen3.8 Max (0902) | Qwen | 1.0M | $2.00 | $6.00 | ā |
| 2026-09-02 | Meta: Muse Spark 1.3 Contributor | Meta | 1.0M | $0.10 | $0.20 | ā |
| 2026-09-02 | Meta: Muse Spark 1.3 | Meta | 1.0M | $1.25 | $4.25 | ā |
| 2026-09-02 | Google: Gemini 3.8 Flash | 1.0M | $0.75 | $3.75 | ā | |
| 2026-09-02 | Google: Gemini 3.8 Flash (batch) | 1.0M | $0.38 | $1.88 | ā | |
| 2026-09-01 | Anthropic: Claude Fable 5.1 | Anthropic | 1.0M | $10.00 | $50.00 | ā |
| 2026-09-01 | Anthropic: Claude Fable 5.1 (batch) | Anthropic | 1.0M | $5.00 | $25.00 | ā |
| 2026-08-31 | IBM: Granite 4.2 8B | ibm-granite | 131K | $0.06 | $0.25 | ā |
| 2026-08-28 | Tencent: Hy4 preview | tencent | 1.0M | $0.83 | $2.50 | ā |
| 2026-08-27 | inclusionAI: Ling 3.0 Flash Fin | inclusionai | 262K | $0.06 | $0.18 | ā |
| 2026-08-27 | inclusionAI: Ling 3.0 Flash Fin (free) | inclusionai | 262K | free | free | ā |
| 2026-08-27 | Z.ai: GLM Flash Latest | z-ai | 1.3M | $0.07 | $0.25 | ā |
| 2026-08-26 | Qwen: Qwen3.8 Flash | Qwen | 1.0M | $0.15 | $0.47 | ā |
| 2026-08-26 | Z.ai: GLM 5.3 Flash | z-ai | 1.3M | $0.09 | $0.30 | ā |
| 2026-08-26 | Z.ai: GLM 5.3 Flash (batch) | z-ai | 1.0M | $0.07 | $0.25 | ā |
| 2026-08-21 | Meta: Muse Spark 1.2 Contributor | Meta | 1.0M | $0.10 | $0.20 | ā |
| 2026-08-21 | DeepSeek: DeepSeek V4 Flash Vision Exp | DeepSeek | 1.0M | $0.22 | $0.66 | ā |
| 2026-08-21 | DeepSeek: DeepSeek V4 Flash Vision Exp (batch) | DeepSeek | 1.0M | $0.11 | $0.33 | ā |
| 2026-08-20 | Tencent: Hy-MT2-1.8B | tencent | 8K | $0.04 | $0.18 | ā |
| 2026-08-20 | Tencent: Hy-MT2-30B-A3B | tencent | 8K | $0.07 | $0.29 | ā |
| 2026-08-19 | Z.ai: GLM Latest | z-ai | 1.3M | $0.77 | $2.43 | ā |
| 2026-08-19 | Tencent: Hy-MT2-7B | tencent | 8K | $0.07 | $0.29 | ā |
| 2026-08-18 | Z.ai: GLM 5.3 | z-ai | 1.3M | $0.91 | $2.86 | ā |
| 2026-08-18 | Z.ai: GLM 5.3 (batch) | z-ai | 1.0M | $0.70 | $2.20 | ā |
| 2026-08-14 | Qwen: Qwen3.8 27B | Qwen | 1.0M | $0.20 | $2.55 | ā |
| 2026-08-14 | Qwen: Qwen3.8 27B (free) | Qwen | 262K | free | free | ā |
| 2026-08-14 | Dots Studio: Dots3-Note Preview (free) | dots-studio | 512K | free | free | ā |
| 2026-08-13 | Google: Gemini 3.7 Flash | 1.0M | $0.75 | $3.75 | ā | |
| 2026-08-13 | Google: Gemini 3.7 Flash (batch) | 1.0M | $0.38 | $1.88 | ā | |
| 2026-08-12 | ByteDance Seed: Seed 2.1 Turbo | bytedance-seed | 262K | $0.50 | $2.50 | ā |
| 2026-08-12 | Qwen: Qwen3.8 2.4T A95B | Qwen | 1.0M | $2.00 | $6.00 | ā |
| 2026-08-12 | Qwen: Qwen3.8 2.4T A95B (batch) | Qwen | 1.0M | $2.00 | $6.00 | ā |
| 2026-08-12 | ByteDance Seed: Seed-2.0-Code | bytedance-seed | 262K | $0.50 | $3.00 | ā |
| 2026-08-12 | DeepSeek: DeepSeek V4 Pro 0813 | DeepSeek | 1.0M | $0.53 | $1.58 | ā |
| 2026-08-12 | DeepSeek: DeepSeek V4 Pro 0813 (batch) | DeepSeek | 1.0M | $0.66 | $1.98 | ā |
| 2026-08-12 | SpaceXAI: Grok 4.6 | xAI | 500K | $2.00 | $6.00 | ā |
| 2026-08-11 | LiquidAI: LFM2.5-2.6B (free) | liquid | 65K | free | free | ā |
| 2026-08-11 | NVIDIA: Nemotron 3.5 Lightning | NVIDIA | 262K | $0.07 | $0.20 | ā |
| 2026-08-11 | NVIDIA: Nemotron 3.5 Lightning (free) | NVIDIA | 1.0M | free | free | ā |
| 2026-08-11 | Sakana: Sakana Namazu | sakana | 262K | $0.95 | $4.00 | ā |
| 2026-08-10 | Upstage: Solar Pro 4 | upstage | 524K | $0.09 | $0.36 | ā |
| 2026-08-09 | Meta: Muse Glimmer 30B | Meta | 131K | $0.30 | $1.20 | ā |
| 2026-08-09 | Meta: Muse Glimmer 30B (batch) | Meta | 131K | $0.17 | $0.75 | ā |
| 2026-08-05 | Meta: Muse Spark 1.2 | Meta | 1.0M | $1.25 | $4.25 | ā |
| 2026-08-01 | DeepSeek: DeepSeek V4 Flash Latest | DeepSeek | 1.3M | $0.04 | $0.16 | ā |
| 2026-07-31 | DeepSeek: DeepSeek V4 Flash 0731 | DeepSeek | 1.3M | $0.04 | $0.16 | ā |
| 2026-07-31 | DeepSeek: DeepSeek V4 Flash 0731 (batch) | DeepSeek | 1.0M | $0.11 | $0.33 | ā |
| 2026-07-30 | Thinking Machines: Inkling Small | thinkingmachines | 1.0M | $0.45 | $1.20 | ā |
| 2026-07-30 | Thinking Machines: Inkling Small (free) | thinkingmachines | 1.0M | free | free | ā |
| 2026-07-27 | Qwen: Qwen3.7 Flash | Qwen | 1.0M | $0.03 | $0.13 | ā |
| 2026-07-24 | Anthropic: Claude Opus 5 | Anthropic | 1.0M | $5.00 | $25.00 | ā |
| 2026-07-24 | Anthropic: Claude Opus 5 (batch) | Anthropic | 1.0M | $2.50 | $12.50 | ā |
| 2026-07-23 | inclusionAI: Ling 3.0 Flash | inclusionai | 262K | $0.02 | $0.06 | ā |
What shipped
Every headline links to the vendor’s own page. Entries marked sitemap come from a vendor that publishes no feed — the page was found by its timestamp and its title read from the page itself.
- OpenAIIntroducing the Australian Youth Safety Blueprint
- GoogleNew experts join Googleās AI & Economy team
- GoogleCo-creating the future of fashion with Google
- AWSKimi K3 by Moonshot AI is now generally available on Amazon Bedrock
- AWSThe new AgentCore Runtime is now available in Amazon Bedrock AgentCore
- AnthropicPartnering with Accenture on embedded evaluationsitemap
- AnthropicHow Claude is uplifting biomolecular modelingsitemap
- xAIIntroducing Grok Voice Transcribe 2.0sitemap
- OpenAIHow Cooley is accelerating IPO work with ChatGPT
- OpenAIIntroducing Astra for Law
- GoogleMaking global data easier to explore
- AnthropicAn alignment assessment of recent cybersecurity incidentssitemap
- AnthropicIntroducing the Life Sciences Verification Programsitemap
- OpenAIHelping older adults use AI in everyday life
- OpenAIReimagining advertising with AI
- OpenAIHex turns complex analysis into visual reports with GPTā6 Astra
- OpenAIHow to connect AI usage to business value
- OpenAIOur framework for reporting model misalignment
- OpenAIHow workers are unlocking new ways of working
- Microsoft AzureMicrosoft recognized as a Leader in the 2026 GartnerĀ® Magic Quadrantā¢ for Distributed Hybrid Infrastructure
- AWSAmazon SageMaker AI now supports serverless model customization for NVIDIA Nemotron 3.5 Lightning
- xAIMemory in Grok Buildsitemap
- GoogleAI for Societal Impact
- GoogleBuilding AI to accelerate science and improve lives
- GoogleAI for everyone in every language
- GoogleNew insights from Googleās AI & Economy ATLAS
- Google DeepMindIntroducing Gemini 3.8 Live and 3.8 Live Extended Thinking
- MicrosoftIntroducing Microsoft 365 G7: Intelligence + Trust for the mission ahead
- Hugging FaceYour Agent Aced the Task. Will It Do It Again?
- AWSAmazon SageMaker AI now supports instance preference lists for training and processing jobs
- AWSAnalyze your CloudTrail events using natural language in Amazon Q Console
- OpenAIHow Fyxer built an AI executive assistant people trust
- OpenAIPerplexity trusts GPT-6 Astra with end-to-end systems
- GoogleWatch astronaut Christina Koch and Googleās James Manyika discuss space, technology, and discovery.
- GoogleDevFest is back
- AWSQwen3.6-35B-A3B-NVFP4 and Wan2.1-T2V-1.3B-Diffusers models now available on Amazon SageMaker JumpStart
- AWSMinistral-3-3B-Instruct-2512 and Ministral-3-8B-Instruct-2512 models now available on Amazon SageMaker JumpStart
- AWSGemma-4-31B-it-assistant and Gemma-4-31B-IT-NVFP4 models now available on Amazon SageMaker JumpStart
- AWSgranite-speech-4.1-2b, kanana-2-30b-a3b-instruct, and OpenFold3 models now available on Amazon SageMaker JumpStart
- OpenAIRapidly scaling online storage to serve over 1 billion ChatGPT users
- OpenAICognition helps Devin test its own work with GPTā6 Astra
- AWSAmazon SageMaker HyperPod now supports model caching for faster inference autoscaling and reduced cold starts
- AWSAmazon Bedrock Managed Knowledge Base now supports multimodal embeddings for video, audio, and image content with TwelveLabs Marengo 3.0
- AWSAmazon SageMaker Unified Studio now supports ODBC connections for Microsoft Power BI and other analytics tools
- AnthropicHow Claude's values vary by model and languagesitemap
- AnthropicEnabling independent research on how people use Claudesitemap
- AnthropicMany-shot jailbreakingsitemap
- AnthropicMapping the mind of a large language modelsitemap
- AnthropicAnthropic Education Report: The AI Fluency Index Ā· Claude Academysitemap
- AnthropicAnthropic Economic Index report: Cadencessitemap
- AnthropicIntroducing the Anthropic Economic Indexsitemap
- AnthropicAnthropic Economic Index: AI's impact on software developmentsitemap
- OpenAIHow a researcher uses Codex and ChatGPT to search for new antimicrobial molecules
- OpenAINow everyone can put data to work
- OpenAIIntroducing ChatGPT for Financial Services
- OpenAIExpanding AI access and cyber defense for federal, state, local, and tribal governments
- OpenAIBuild more natural voice experiences with GPTāLiveā1 in the API
- OpenAIIntroducing the Agents API
- Google3 ways to prep for your next big race with Search
- MicrosoftESPC 2026 is coming to Amsterdam: Bringing Microsoft and the community together
- MicrosoftBuild apps in Copilot Cowork and Copilot Studio
- Microsoft AzureMicrosoft named a Leader in the 2026 GartnerĀ® Magic Quadrantā¢ for Container Management
- Microsoft AzureThe Economics of Agent Optimization: How AI agent governance controls cost and proves ROI
- Microsoft AzureThe future of infrastructure resiliency starts with modernization
- Hugging FaceAsync GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL
- Hugging FaceRebuilding AUTOMATIC1111 with Gradio Workflow
- OpenAIThe AI policy window is open. We need to act.
- OpenAIGPT-6 Astra: The next generation in intelligence for work
- OpenAIPaul Christiano joins OpenAI Foundation Board
- GoogleGet ready for the game with new football features in Search
- GoogleRecreating a 70-year love story frame by frame
- Microsoft AzureTwo zones or three? A design framework for zone-resilient Azure workloads
- AWSAmazon Bedrock Managed Knowledge Base adds APIs and console support for debugging document-level access control
- AWSAmazon Bedrock Managed Knowledge Base now supports Confluence Data Center as a native data source connector
- OpenAIHow GPT-5.6 Sol helps run quantum computing experiments
- OpenAIThe Work Now Within Reach
- OpenAIIntroducing ChatGPT Images 2.5
- OpenAIOn the NavierāStokes Millennium Prize Problem
- OpenAIFunding grants for new research into AI and teen development
- OpenAIOpenAI expands initiatives to support journalism from classrooms to newsrooms
- OpenAI1Password increases engineering productivity 21% with Codex
- Google DeepMindAlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome
- Microsoft AzureBeyond the benchmark: How an adaptive approach drives scientific discovery
- Hugging FaceSafety for Whom? Refusing the Right Subset of a Topic, Not the Whole Topic
- AWSOpenAI GPT-6 Astra is now generally available on Amazon Bedrock
- AWSAmazon SageMaker Feature Store now supports individual feature updates to lower write latency
- AWSAmazon Bedrock AgentCore Memory now supports direct ingestion to long-term memory
- OpenAISupporting independent journalism in Ukraine
- OpenAIAn Alien Mind
- OpenAIResearch acceleration: The view inside OpenAI
- xAISetting Grok Bot loose on procurementsitemap
- OpenAIDaybreak for Frontline Defenders: $1B to protect essential services
- OpenAILegora reviewed 41 documents in minutes with GPT-6 Astra
- OpenAIPlayco cut manual fixes 50% prototyping games with GPT-6 Astra
- OpenAIGPT-6 Astra: A new generation of intelligence
- OpenAISafety overview: GPT-6 Astra
- Google DeepMindIntroducing WeatherNext 3, our most advanced and accurate global weather AI model
- MicrosoftPPCC 2026: Bringing AI and your business processes together
- Microsoft AzureEnterprise AI transformation relies on the end-to-end platform: Azure was built for this moment
- Microsoft AzureGPT-6 Astra: Frontier intelligence for work, now generally available in Microsoft Foundry
- Microsoft AzureHow Microsoft’s Physical Security Engineering Team scaled hybrid operations with Azure Arc and Azure Virtual Desktop
- Hugging FaceNeoMME: an efficient Multimodal-native and Multilingual Encoder
- Hugging FaceFine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
- Hugging FaceGive Your Coding Agents a Memory You Own
- Hugging FaceTraining a coding model to paint watercolours with TRL and OpenEnv
- xAIDesigning Grok Bot for a world of persistent agentssitemap
- xAIGrok Bot for Enterprisesitemap
- OpenAIATV Big Air Tour turned 3 days of work into 3 hours with ChatGPT
- GoogleProactive cyber defense for governments and enterprises
- Google DeepMindProactive cyber defense for governments and enterprises
- Google DeepMindIntroducing Gemini 3.8 Flash and 3.8 Flash Cyber
- Microsoft AzureThe Economics of Agent Optimization: Context engineering for enterprise AI agents
- OpenAIHow AI-native companies turn workflows into operating capability
- OpenAIPath to Astra: critical capabilities and frontier safeguards
- OpenAIHealthcare organizations can now connect EHR and additional industry data to ChatGPT
- OpenAIHow law firm Gilbert + Tobin governs and scales AI with OpenAI
- GoogleThe latest AI news we announced in August 2026
- GoogleTry Google Pics: Easy image creation and editing in Google Workspace
- Google DeepMindIntroducing agentic video understanding with Gemini
- Hugging FaceBenchMIRT: What are LLM benchmarks actually measuring?
- Hugging FaceIntroducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI
- xAIBiosecurity at the frontiersitemap
- OpenAIOpenAI supports Californiaās bill to advance youth AI safety
- OpenAIPolimill builds Japan's next-generation public AI infrastructure
- OpenAIA milestone in expanding access to AI
- xAIGrok Bot now works with Xsitemap
- OpenAIOur decision on Cursor following its acquisition by SpaceX
- OpenAISupporting Thailandās next generation of AI startups
- Hugging FaceThe Open ASR Leaderboard Adds Its First Global South Language
- OpenAIBetter answers, broader thinking: What students gain from ChatGPT and critical-thinking training
- OpenAIExpanding OpenAIās presence in Brazil
- Google3 new ways to plan and book travel in Search
- Google DeepMindGemini Omni 1.1 Flash lets you build with more control
- Google DeepMindPiloting the world's first double-blind AI evaluations
- OpenAIBringing ChatGPT for Teachers to more U.S. school districts
- OpenAILearning never stops: How AI makes learning continuous
- OpenAIThe Hugging Face incident and the road ahead
- OpenAIHow loveholidays is making everyone a builder with Codex
- Google DeepMindIntelligent transcription with Gemini 3.5 Transcribe
- Hugging FaceTraining and Finetuning Multi-Vector Embedding Models with Sentence Transformers
- xAIGrok 4.6 on Microsoft Foundrysitemap
- xAIGrok Bot is now included with more planssitemap
- OpenAIThe full stack behind abundant intelligence
- OpenAIJalapeƱoās first results show industry-leading speed and efficiency in AI inference
- OpenAIDisrupting a new covert influence campaign from Russia
- OpenAIIntroducing the Admin plugin for ChatGPT Work and Codex
- Google5 ways to upgrade your home decor with Google Search
- Hugging FaceGranite 4.2 LLMs: How They're Built
- Hugging FaceQuantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original
- Hugging FaceWire It, Run It, Deploy It: AI Workflows in Gradio
- OpenAIAdvancing price-performance for developers with GPTā5.6 in Kiro
- Google DeepMindFrom Atari to EVE Online: Building on 15 Years of AI Research in Games
- Hugging FaceHow Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code
- Hugging FaceMeasuring benchmark optimization in speech recognition
- xAIGrok 4.6 on Gemini Enterprise Agent Platformsitemap
- OpenAIIntroducing Intelligence Age
- OpenAIStampli cuts launch hours by 68% using ChatGPT Work
- Hugging FaceUp to 3.2x Faster Inference with LFM2.5-DSpark
- OpenAIOffering Zero Data Retention for frontier models
- OpenAIReplit expands access to software creation with GPT-5.6 Luna
- Google5 new ways to level up your learning with Search
- xAIGrok 4.6 on Amazon Bedrocksitemap
- xAIGrok Build on web and mobilesitemap
- OpenAIChatGPT Ads expands across Europe
- OpenAIStrengthening democratic oversight in national security
- OpenAIPartnering with CodeAI to prepare the first AI generation
- OpenAIPacing model development in an era of cyber-critical capabilities
- OpenAIIntroducing ChatGPT for Teens: Built for learning, backed by protections
- OpenAIHow NVIDIA scales expertise with ChatGPT Work
- OpenAIAsana cleared 5 years of engineering work in 2 weeks with Codex
- Hugging FaceHow Much Memory Does Your Agent Actually Need?
- Hugging FaceMulti-Vector (Late Interaction) Embedding Models with Sentence Transformers
- OpenAIThe Defenderās Window
- OpenAIOpenAI joins PORTS-Pike project
- OpenAINew policy ideas for the Intelligence Age
- GoogleGet closer to the game with Gemini and Pixel
- Hugging FaceSame Cluster, 33 Points More Utilization: What Changed Was the Order
- Hugging FaceState of Open Models: Summer 2026 Observations
- OpenAIThe builderās guide to GPTā5.6
- OpenAIPreviewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed
- OpenAIOpenAI appoints Dali Rajic as Chief Revenue Officer
- GoogleBring your spreadsheet data to life with Sheets canvas
- Google DeepMindIntroducing Gemini 3.7 Flash
- Hugging FaceRecord, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets
- Hugging FaceWhat We Learned by Reproducing 2,200 papers from ICML
- OpenAIFrom assistance to execution: How enterprises put AI to work
- OpenAIHow RingCentral builds AI-native work from engineering to ops
- Google DeepMindPutting sign language AI into usersā hands
- Hugging FaceIntroducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis
- OpenAIDaybreak models are now available on AWS
- OpenAITesting ads in ChatGPT
- Hugging FaceThinking of ACE? We Can Do It with Fewer Tokens
- OpenAIOpenAIās letter to Governor Abbott on responsible AI infrastructure in Texas
- OpenAIModel ML completes finance work more efficiently with GPT-5.6 Sol
- OpenAIWhat building an AI-native finance function taught me
- OpenAIPutting frontier cyber models in more trusted hands
- OpenAIExpanding Daybreak as the Cyber Defense Window Narrows
- OpenAIVirgin Atlantic sharpens customer journeys with ChatGPT Work
- OpenAIHow Zapier transformed core marketing processes with ChatGPT Work
- OpenAIPremium seats are coming to ChatGPT Business
- Hugging FaceBuild Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS
- Hugging FaceMaking Knowledge Distillation Cheap Enough to Run at Scale
- Hugging FaceMeta is back with Muse Glimmer: local, agentic, multimodal, and open source
- OpenAIResponding to the next frontier of critical cyber capabilities
- OpenAIHow HSP GRUPPE builds AI capabilities for tax advisory
- OpenAIImproving GPTā5.6 Sol in ChatGPTāand expanding access to GPT-5.6 Luna for free users
- OpenAIWorking with the American Psychological Association on youth mental health and AI
- OpenAIFrom asking to doing: How the world is putting ChatGPT to work
- Google DeepMindWeatherNext: AI model achieves breakthrough in forecasting cyclones
- Hugging FaceBaseten on Hugging Face Inference Providers š„
- OpenAIThird-party cyber evaluations involving OpenAI models
- OpenAINew ways to learn and teach with ChatGPT Work and Codex
- OpenAIApple is getting this wrong
- OpenAIHow we built a realtime system for responsive voice AI in six months
- OpenAICircles powers telco personalization with OpenAI technology
- OpenAITen advances in mathematics and theoretical computer science
- OpenAIAdvancing responsible AI across Europe
- OpenAIBuilding abundant intelligence
- OpenAIUnivƩ builds an AI-ready workforce
- OpenAIDisrupting a Cambodia-based criminal scam operation
- OpenAIAdvancing the price-performance frontier with GPT-5.6
- OpenAIHow avatarin built a 24/7 retail agent with GPT-Realtime
- Google DeepMindGemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration
- MicrosoftThe next measure of AI momentum is work transformed
- Hugging FaceGPU Management: Why Idle GPUs Are the New Grounded Aircraft
- OpenAIHow enabling two settings tripled our scores on the ARC-AGI-3 benchmark
- OpenAIAccelerating scientific discovery with ChatGPT for Academic Researchers
- OpenAIHow GPT-5.6 fuses frontier intelligence with frontier efficiency
- Google DeepMindWeāre launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
- OpenAIScientific computing in the age of agentic AI
- Google DeepMindGemini Robotics 2 brings whole body intelligence to robots
- OpenAIHow AI is expanding what people do at work
- Hugging FaceNVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics
- Hugging FaceAnatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
- OpenAILaunching Health in ChatGPT
- Hugging FaceBringing Nunchaku 4-bit Diffusion Inference to Diffusers
- OpenAIBuilding AI infrastructure with the Effingham County community
- OpenAIHow news organizations are using AI to advance their vital missions
- OpenAIAdvancing the next era of national science
- OpenAIIntroducing OpenAI Presence
- OpenAINTT DATA Group cuts incident analysis to 30 minutes with Codex
- Google DeepMindAccelerating the frontiers of scientific discovery: Googleās $40M commitment to the Genesis Mission
- OpenAIIntroducing the ChatGPT for small business program
- OpenAIOpenAI and Hugging Face partner to address security incident during model evaluation
- OpenAIDavid VĆ©lez and Robin Vince join the boards of the OpenAI Foundation and OpenAI Group PBC
- Google DeepMindIntroducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
- Hugging FaceGrabette: an open system to record robot-manipulation data
- OpenAISafety and alignment in an era of long-horizon models
- OpenAIA scorecard for the AI age
- Google DeepMindIntroducing Gemini 3.5 Flash Cyber
- OpenAIWhy teens deserve access to safe AI
- OpenAIHow Codex became a collaborator for OpenAIās creative team
- OpenAIHow Cars24 scales conversations and builds faster with OpenAI
- Google DeepMindOur approach to bioresilience
- Hugging FaceNewer Models, Same Advantage
- Hugging FaceSecurity incident disclosure ā July 2026
- OpenAIThe US is advancing AI safety through state and federal action
- OpenAIGPT-Red: Unlocking Self-Improvement for Robustness
- Hugging FaceModel Routing Is Simple. Until It Isnāt.
- Hugging FaceWelcome Inkling by Thinking Machines
- Hugging FaceIntroducing Real World VoiceEQ: Measuring the human quality of voice AI
- OpenAIHow to manage AI investments in the agentic era
- OpenAIData science workflows with ChatGPT Work
- OpenAISales workflows with ChatGPT Work
- Google DeepMindEmpowering Indiaās next generation of innovators with ATL Saathi
- OpenAIHow Deutsche Telekom is rewiring telecommunications with AI
- OpenAIGetting started with ChatGPT
- Hugging FaceProfiling in PyTorch (Part 3): Attention is all you profile
- OpenAIGPT-5.6 is now the preferred model in Microsoft 365 Copilot
- OpenAIGPT-5.6: Frontier intelligence that scales with your ambition
- OpenAIGPT-5.5 Bio Bug Bounty
- OpenAIChatGPT is now a partner for your most ambitious work
- OpenAIOur approach to government and national security partnerships
- OpenAISeparating signal from noise in coding evaluations
- OpenAIHelping Kā12 educators build practical AI skills
- OpenAIIntroducing GPT-Live
- Hugging FaceNative-speed vLLM transformers modeling backend
- OpenAIAustralian Payments Plus moves faster with ChatGPT and Codex
- OpenAIMUFG aims to become AI-native with OpenAI
- Hugging FaceFrom Hugging Face to Amazon SageMaker Studio in one click
- Hugging FaceHugging Face Models on Foundry Managed Compute
- Hugging FaceRun AI workloads on any cloud, store on Hugging Face: zero-egress storage with SkyPilot
- Hugging FaceLeRobot v0.6.0: Imagine, Evaluate, Improve
- Hugging FacePRX Part 4: Our Data Strategy
- Hugging Faceš¤ Kernels: Major Updates
- Google DeepMindGoogle DeepMind and A24 announce first-of-its-kind research partnership
- Hugging FaceHugging Face and Cerebras bring Gemma 4 to real-time voice AI
- OpenAIHow ChatGPT adoption has expanded
- OpenAICore dump epidemiology: fixing an 18-year-old bug
- OpenAIInside Genebench-Pro
- OpenAIIntroducing GeneBench-Pro
- Google DeepMindStart building with Nano Banana 2 Lite and Gemini Omni Flash
- Hugging FaceScarfBench: Benchmarking AI Agents for Enterprise Java Framework Migration
- Hugging FaceWhy Specialization Is Inevitable
- Hugging FaceFeaturing Every Eval Ever Results on Hugging Face Model Pages
- OpenAIMapping Europeās AI Workforce Opportunity
- Hugging FaceDiScoFormer: One transformer for density and score, across distributions
- OpenAIHP Inc. launches Frontier strategic partnership with OpenAI
- OpenAIPreviewing GPT-5.6 Sol: a next-generation model
- Hugging FaceRun a vLLM Server on HF Jobs in One Command
- OpenAIHow agents are transforming work
- MicrosoftCopilot in Excel: Built for the era of Frontier Finance
- OpenAIOpenAI and Broadcom unveil LLM-optimized inference chip
- Google DeepMindIntroducing computer use in Gemini 3.5 Flash
- Hugging FaceAccelerating Transformers Fine-Tuning with NVIDIA NeMo AutoModel
- Hugging FaceIntroducing the FFASR Leaderboard: Benchmarking ASR in the Real World
- OpenAIHelping build shared standards for advanced AI
- OpenAIHow GPT-5 helped immunologist Derya Unutmaz solve a 3-year-old mystery
- OpenAIHow Omio is building the future of conversational travel
- Hugging FaceShipping huggingface_hub every week with AI, open tools, and a human in the loop
- Hugging FaceExperimenting with the proposed Cross-Origin Storage API in Transformers.js
- OpenAIPatch the Planet: a Daybreak initiative to support open source maintainers
- OpenAIDaybreak: Tools for securing every organization in the world
- OpenAICodex-maxxing for long-running work
- Hugging FacePP-OCRv6 on Hugging Face: 50-Language OCR from 1.5M to 34.5M Parameters
- Hugging FaceWe got local models to triage the OpenClaw repo for FREE!*
- OpenAISamsung Electronics brings ChatGPT and Codex to employees
- OpenAINew usage analytics and updated spend controls for enterprises
- OpenAIImproving health intelligence in ChatGPT
- OpenAIUsing AI to help physicians diagnose rare genetic diseases affecting children
- Hugging FaceMosaicLeaks: Can your research agent keep a secret?
- Hugging FaceBeyond LoRA: Can you beat the most popular fine-tuning technique?
- Hugging FaceIs it agentic enough? Benchmarking open models on your own tooling
- OpenAIA near-autonomous AI chemist improves a challenging reaction in medicinal chemistry
- OpenAIIntroducing LifeSciBench
- Hugging FaceFrom the Hugging Face Hub to robot hardware with Strands Agents and LeRobot
- Hugging FaceGLM-5.2: Built for Long-Horizon Tasks
- Hugging FaceAgentic Resource Discovery: Let agents search
- OpenAIPredicting model behavior before release by simulating deployment
- Google DeepMindUnlocking UK house-building with AI-accelerated planning
- Google DeepMindSecuring the future of AI agents
- MicrosoftCopilot Cowork is now generally available
- OpenAIIntroducing the OpenAI Partner Network
- OpenAINew OpenAI Academy courses for the next era of work
- OpenAIHow Preply combines AI and human tutors to personalize learning
- OpenAIHow an astrophysicist uses Codex to help simulate black holes
- OpenAIOpenAI to acquire Ona
- OpenAISupporting Europeās work in ensuring a trustworthy AI ecosystem
- OpenAIBBVA puts AI at the core of banking with OpenAI
- Hugging FaceProfiling in PyTorch (Part 2): From nn.Linear to a Fused MLP
- OpenAIAccess OpenAI models and Codex through your Oracle cloud commitment
- OpenAIPRC-linked influence operations are targeting AI debates in the US
- OpenAIFrom data to decisions: how LSEG is scaling trusted AI
- Google DeepMindDiffusionGemma: 4x faster text generation
- Google DeepMindInvesting in multi-agent AI safety research
- OpenAIHow engineers at Nextdoor use Codex to build without limits
- OpenAIWhat Codex unlocks for Notion
- OpenAIIndustrial policy for the Intelligence Age
- Google DeepMindFluid, natural voice translation with Gemini 3.5 Live Translate
- Google DeepMindIntroducing Gemma 4 12B: a unified, encoder-free multimodal model
- Google DeepMindPowering the future of robotics in Europe
- Hugging FaceHow an Agent Built a 3D Paris Gallery by Chaining Two Hugging Face Spaces
- Hugging FaceMigrating Your GitHub CI to Hugging Face Jobs
- OpenAIConfidential submission of draft S-1 to the SEC
- OpenAIBuilt to benefit everyone: our plan
- OpenAIIntroducing the OpenAI Economic Research Exchange
- Google DeepMindMeasuring the impact of learning with AI in Sierra Leone and beyond
- Hugging FaceThe Open Source Community is backing OpenEnv for Agentic RL
- OpenAIHow Endava is redesigning software delivery around AI agents
- OpenAIDreaming: Better memory for a more helpful ChatGPT
- OpenAIBiodefense in the Intelligence Age
- Hugging FaceNemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI
- Hugging FaceDesigning the hf CLI as an agent-optimized way to work with the Hub
- OpenAIIntroducing new capabilities to GPT-Rosalind
- OpenAIHow Wasmer used Codex to build a Node.js runtime for the edge
- OpenAIA blueprint for democratic governance of frontier AI
- OpenAIOpenAI public policy agenda
- Hugging FaceDirect Preference Optimization Beyond Chatbots
- Hugging FaceAdding MCP Tools to Reachy Mini
- OpenAITravelers deploys AI-powered claims countrywide with OpenAI
- OpenAIChatGPT Sites
- OpenAICodex for every role, tool, and workflow
- OpenAIAdvancing youth safety and opportunity through global leadership
- OpenAICodex is becoming a productivity tool for everyone
- MicrosoftIntroducing Microsoft Scout: Your always-on personal agent
- MicrosoftAnnouncing the new Work IQ APIs
- Hugging FaceHolo3.1: Fast & Local Computer Use Agents
- OpenAIOur views on AI policy and political advocacy
- OpenAIBuilding the infrastructure for the Intelligence Age in Michigan
- OpenAIOpenAI frontier models and Codex are now available on AWS
- OpenAIāData Center Bandwagonā Campaign: US-targeted influence activity
- OpenAIāTech and Tariffsā Campaign: Influence activity targeting US tech policy
- Hugging FaceIntroducing Mellum2: A 12B Mixture-of-Experts Model by JetBrains
- Hugging FaceBeyond LLMs: Why Scalable Enterprise AI Adoption Depends on Agent Logic
- OpenAIBoston Childrenās uses AI to unlock new diagnoses
- OpenAIHow Braintrust turns customer requests into code with Codex
- OpenAIStrengthening societal resilience with Rosalind Biodefense
- OpenAIA shared playbook for trustworthy third party evaluations
- Hugging FaceProfiling in PyTorch (Part 1): A Beginner's Guide to torch.profiler
- OpenAIHow Endava builds an agentic organization with Codex
- OpenAIOpenAIās Frontier Governance Framework
- MicrosoftIntroducing Microsoft 365 Business with Copilot: The new standard for small business
- OpenAICisco and OpenAI redefine enterprise engineering with Codex
- OpenAIBuilding self-improving tax agents with Codex
- OpenAIWarpās big bet on building open source with GPT-5.5
- OpenAIElection information and safeguards in 2026
- Hugging FaceReachy Mini goes fully local
- Hugging FaceShipping a Trillion Parameters With a Hub Bucket: Delta Weight Sync in TRL
- OpenAIAI is becoming a first hire for small businesses
- OpenAIOpenAI, Grupo Folha and Grupo UOL announce strategic content partnership
- Hugging FaceHarness, Scaffold, and the AI Agent Terms Worth Getting Right
Nothing matches those filters.
Where this comes from
A job on a clock reads these once a day and commits what it finds, so the page is current whether or not I open a laptop. Last refreshed . The store only grows: Google’s AI feed holds 20 items and Microsoft’s 10, so anything not written down within a fortnight would be gone for good.
| Vendor | How | Source |
|---|---|---|
| OpenAI | their RSS feed | openai.com/news/rss.xml |
| their RSS feed | blog.google/technology/ai/rss/ | |
| Google DeepMind | their RSS feed | deepmind.google/blog/rss.xml |
| Microsoft | their RSS feed | www.microsoft.com/en-us/microsoft-365/blog/feed/ |
| Microsoft Azure | their RSS feed | azure.microsoft.com/en-us/blog/feed/ |
| Hugging Face | their RSS feed | huggingface.co/blog/feed.xml |
| AWS | their RSS feed | aws.amazon.com/about-aws/whats-new/recent/feed/ |
| Anthropic | sitemap timestamps (no feed published) | www.anthropic.com/sitemap.xml |
| xAI | sitemap timestamps (no feed published) | x.ai/sitemap.xml |
| OpenRouter | model catalogue API | openrouter.ai/api/v1/models |
The catalogue is OpenRouter’s, which is an aggregator rather than each vendor’s own API: it lists community models beside official ones and its prices are its own. It is used because it is the one source that answers for every vendor at once — three of them publish no model API at all.