Model intelligence.
Follow the changes. Understand what comes next.
Google Launches Gemini 4 Neon in Antigravity With 1M Context, Four Thinking Levels, and Benchmarks vs. Claude Opus 5.5 and GPT-6.2 Sol
Announced on November 19, 2026 at 05:51 Beijing time, Google's Gemini 4 Neon features a 1M-token context window, four thinking levels, and planner-executor pairing with Gemini 4 Hydrogen in Google Antigravity, outperforming Claude Opus 5.5 and GPT-6.2 Sol across most of 18 published benchmarks.
Latest stories349
Newest publishedMoonshot AI Launches High-Speed Kimi K3.1 Pro Prime on Alibaba Model Studio: 1.5x–2x TPS with 1M Context Intact
Listed on November 19, 2026 on Alibaba Cloud Model Studio's Prime Mode (and tracked across global gateways on November 20), kimi-k3.1-pro-prime is the high-speed Prime variant of Moonshot AI's 3T-class Kimi K3.1 Pro. It keeps identical reasoning, vision, tool-calling, and 1M-token capabilities while boosting output TPS by 1.5x–2x at RMB 26 input / RMB 136 output / RMB 1.3 cached hit per 1M tokens in Beijing.
ChatGPT Space Rolls Out Globally to Plus Users Ages 18+, Separate from Dot's Regional Limits and Slides Rollout
On November 20, 2026, OpenAI announced in its ChatGPT release notes that ChatGPT Space is rolling out globally to Plus subscribers ages 18 and older. Independent from the previous day's region-restricted Dot test, Space carries no EEA/UK/Switzerland exclusion for Plus—while Slides rolls out separately to eligible Space accounts, meaning Space access does not automatically unlock Slides.
Thinking Machines Lab Open-Sources Inkling 2 Mini: A 26.4B Looped MoE Model Running Offline in 20 GB
On November 20, 2026, Thinking Machines Lab released Inkling 2 Mini under the Apache 2.0 license—a multimodal looped MoE transformer (26.4B total, 3.4B base active, 5.4B effective active across 3 middle-layer loops) with a 1M context window. Trained and deployed natively in mixed NVFP4 and MXFP8 (approx. 20 GB single checkpoint), it runs locally on 32 GB+ Apple Silicon Macs or a single NVIDIA DGX Spark while leading sub-100B open-weight coding benchmarks.
Tencent Opens Flagship Hy4 Weights (770B Total / 49B Active): Variant Selection Guide and Unity, Unreal, and Godot Adaptation
On November 20, 2026, Tencent opened weights for its flagship multimodal MoE model Hy4 (originally announced November 4, distinct from the lighter 220B Hy4 Flash line) across Hugging Face, ModelScope, and GitCode in three variants: tencent/Hy4, tencent/Hy4-FP8, and tencent/Hy4-base. Featuring a 770B language backbone with 49B activated parameters per token and 1M context, Hy4 adds dedicated adaptation for Unity, Unreal Engine, and Godot.
Google DeepMind Opens Genie 4 World Model to Google AI Ultra Subscribers and Launches GA on Gemini Interactions API
Between November 18 and 19, 2026, Google DeepMind ended the closed research preview for Genie 4—opening Project Genie to eligible Google AI Ultra subscribers (with a waitlist for Pro users) and launching models/genie-4 to general availability on the Gemini Interactions API (v1beta). Built on Gemini Omni Flash and Gemini reasoning, Genie 4 streams real-time 720p 24 fps interactive worlds with 48 kHz stereo audio at $17.50 per 1M video output tokens (approx. $0.10/sec).
OpenAI Launches ChatGPT for Accounting with Dedicated Accounting Plugin and QuickBooks, Xero, and NetSuite Integrations
Announced on November 18, 2026 (early November 19 Asia/Shanghai), OpenAI introduced ChatGPT for Accounting—combining its dedicated Accounting plugin and optional Computer Use with partner integrations including Intuit QuickBooks, Xero, BILL, FreshBooks, NetSuite, QuickStart, Rillet, and Campfire to streamline invoice coding, bank reconciliation, and month-end reporting while keeping final approval in human hands.
OpenAI Expands ChatGPT dots to Select Plus Users in Small Test and Rolls Out to Business Standard
On November 19, 2026 (07:17 Asia/Shanghai on November 20), OpenAI's Tibo (@thsottiaux) announced a small-scale test of ChatGPT dots for ChatGPT Plus users alongside a rollout for ChatGPT Business Standard. Per OpenAI's updated Help Center docs, Plus testing excludes the EEA, UK, and Switzerland, Business Standard rollout currently targets eligible U.S. members (one personal dot per member), and iMessage/RCS texting remains limited to select Pro users.
Gemini App for Mac Gains Background Computer Use in an Isolated Window with App-Level Permissions
On November 19, 2026, Google introduced Computer Use in the Gemini app for Mac, allowing Gemini to navigate desktop apps, local files, and websites inside an isolated background virtual screen without taking over your cursor. Rolling out in English in the U.S. on Apple Silicon Macs running macOS 15.0+, the feature lists Google AI Plus, Pro, and Ultra on the launch blog (now aligned with the updated Help Center page after earlier help text listed only Pro and Ultra).
Google Launches Gemini 4 Neon in Antigravity With 1M Context, Four Thinking Levels, and Benchmarks vs. Claude Opus 5.5 and GPT-6.2 Sol
Announced on November 19, 2026 at 05:51 Beijing time, Google's Gemini 4 Neon features a 1M-token context window, four thinking levels, and planner-executor pairing with Gemini 4 Hydrogen in Google Antigravity, outperforming Claude Opus 5.5 and GPT-6.2 Sol across most of 18 published benchmarks.
Google DeepMind World Model Genie 4 Launches in Project Genie With Mid-Session Text Edits, Real-Time Audio, and Genie 3 Upgrades
On November 18, 2026 (November 19, 2026 at 02:15 Beijing time), Google announced that Google DeepMind's Genie 4 world model is now available in Project Genie with mid-session environment editing and native audio, alongside Project Genie agent, a community gallery, and 3-minute sessions for Genie 3.
Xingchen AGI Lab Publishes TeleChat Xing 4 Technical Report and TeleGym Platform Trained on 8,192 Ascend 910C Chips
Published on arXiv on November 16, 2026 (arXiv:2611.22115, featured #1 on Hugging Face Daily Papers on November 18, 2026), Xingchen AGI Lab's report 'TeleChat to Xing4' details pretraining a 285B-parameter agentic model across 8,192 Huawei Ascend 910C accelerators using MindSpore HyperParallel and introduces TeleGym for asynchronous agentic post-training.
Antigravity Can Now Control Robots: Google Launches the Robotics Plugin With MuJoCo Simulation and Real-World Pick-and-Place
On November 17, 2026, Google introduced the Robotics plugin for Google Antigravity. Powered by Gemini 4 Hydrogen, the plugin integrates USB hardware bring-up, in-IDE MuJoCo physics simulation, spatial perception, and closed-loop robot control—demonstrated in an end-to-end workflow from a whiteboard sketch to a physical arm sorting snacks off a conveyor belt.