China just

MINIMAX Built for the second look. We weave a quiet intelligence layer, into Beating top AI Models SVG-Bench Terminal Bench 2.1 GDPval rubrics GRIND BUILT IN SILENCE CHAPTER OF A BONFICE HEMEL PRODUCTION 0 Here's 45 that happened in in under 60 seconds Google 6 Google AI Edge Eloquent Usage stats 12768 words Total number of words dictated. Cleaned up text Yeah, let's meet for dinner on Thursday at 6:30. Is this what you meant to say? Original text 0 inserts, 0 deletes Polish Key points Formal Short Listening... Live transcription Hey, uh so we would like you to review the new tagline for the Google AI Edge eloquent app. Uh this was actually proposed by Morgan and it's already been approved by Sir Rio and Amy. It goes speak naturally, express collectively no cap. This is fun because no cap is colloquial, it means, you know, truthfully, but in this case it also means there's no app on the no cap on the usage of the app and because it is capable of running on device and because there is by own device Polishing transcript... Calculating stats for this session... 93 words 186 wpm Stats from your last dictation. Cleaned up text Hey, so we'd like you to review the new tagline for the Google AI Edge Eloquent app. This was proposed by Mogan and already approved by Suril and Ami. It goes: "Speak naturally, express effectively, no cap." This is fun because "no cap" is colloquial, meaning truthfully, but in this case, it also means there's no app on the no cap on the usage of the app and because it's capable of running on device and it requires no server subscription. 20 inserts, 50 deletes Polish Key points Formal Short Polished text copied to clipboard! Record your ideas, Eloquent polishes them instantly. RAW INPUT POLISHED Um can we push the project review with uh Surreal to next Tuesday? And uh also ask Amy to Wednesday as he might still be traveling. And uh also ask Amy to send over the latest uh feedback doc. I mean the user feedback doc. So Morgan can include it in the in the uh presentation as well. cleans your speech in real time and never touches the cloud 5 Rever 2.0 Best 4K Best 4K Image Model that lets you grab and adjust free to try and already Image Arena Ideogram 4.0 Gemini 3.1 flash Gemini 3 Pro Image Preview w 2k (nano-banana-pro) Gemini 3 Pro Image Preview w (nano-banana-pro) GPT-image-2 (medium) GPT-image-1.5 high-fidelity GPT 5.5 Gemini 3.1 Pro Claude Sonnet 4.6 Claude Opus 4.8 Nemotron 3 Ultra Max NVIDIA AI @NVIDIAAI Follow Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model built for long-running agents. It delivers 5x faster inference and lowers the cost of complex agentic tasks by up to 30% versus other open frontier models. especially for a single user's inference trace. a 550 billion open model As always, Nemotron 3 Ultra is fully open. This includes model weights, synthetic data, and post-training recipes. Available now on @huggingface -> nvidia /NVIDIA-Nemotron-3-Ultra-550B- A55B-NVFP4 nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4... From huggingface.co open weights live on Hugging Face 3 Ideogram Ideogram 4.0 an incredible text to image model with flawless native 2k output and JSON layout control Now ranked it top one on the open weight design arena Microsoft 2 Meet Scout and always on Microsoft FY2022 Budgeting Scout Sending you the pre-read for the FY2027 budget review meeting FY2027 Budget Review John Kirkland Today, 8:10 AM Thank you Scout, will review! Working on it Hey, Scout, can you keep an eye on my inbox and Teams for any outstanding decisions that are waiting on me? Absolutely. I'll monitor your inbox and Teams, and let you know if anything important comes up. Automations Morning briefing Teams monitoring Outlook monitoring Integrations Teams Outlook your habits your meetings runs your inbox 24/7 across and web 1 Minimax M3 Minimax M3 M3 M3 Minimax M3 Intelligent Frontier. 1M-context MSA. Native Multimodality API & Token Plan weight model Try in Minimax Code Read the Report frontier 1M Tokens a 1 million token context make it M3's championship run Analyzing Analyzing reference... scroll-driven timeline, 9 phases Analyzing brand asset... extracted palette, type, voice Building Coding scroll-driven layout (GSAP ScrollTrigger) and native image input in one architecture Beating M3 Opus 4.7 GPT 5.5 Gemini 3.1 Pro 63.7 62.3 58.2 59.2 SVG-Bench GPT 5.5 78.2 66.0 66.1 70.0 Terminal Bench 2.1 77 74.7 70.8 57.8 GDPval rubrics MCP Atlas Pluss it even beats GPT 5.5 on the coding benchmarks to learn about and tutorials follow