← AM·PM Brief

🚀 Hy3 Hits #1 on OpenRouter, DeepMind Launches Gemini Embed 2 for Search - AI Daily Brief

· Evening brief · 12 news · 8:20

Audio in Mandarin Chinese · English transcript below

MystModel,AppleGoogle,NewClaude,$5BInv,AGI,Multi,EffRev,EntTrans,TechFst,EvalUp

Today's Top 3 Headlines

  1. AI Industry News

    🤖 Hy3 LLM tops OpenRouter model leaderboard by a huge margin

    Minimaxir.com reveals that a mysterious LLM named Hy3 tops OpenRouter leaderboard by a significant margin, surpassing models like Claude and even generating complex HTML code. For the AI industry, this highlights transparency, data governance, and trust challenges as LLMs pursue performance, developers…

    Source
  2. AI News

    🤖 Apple integrates Google's trillion-parameter Gemini model into Siri

    Apple plans to integrate Google's trillion-parameter Gemini model into iPhones, with the new Siri running a distilled on-device model combined with cloud-based Google infrastructure, per The Information. For Apple, this marks a shift from purely on-device to hybrid AI strategy, with user privacy protected via NVIDIA confidential computing.

    Source
  3. AI Product Launch

    🚀 Anthropic releases Claude Opus 4.8 with dynamic workflows, up to 1000 sub-agents

    Anthropic launches Claude Opus 4.8 with dynamic workflows, cheaper fast mode, and up to 1000 sub-agents. For AI product developers, this means building more complex multi-step automations while reducing costs for high-frequency calls.

    Source

+8 more headlines

  • 🤖 AutoTTS automates LLM inference strategy, reduces token usage by 69.5%
  • 🛡️ IBM and Red Hat pledge $5B for AI-driven open-source security initiative
  • 🤖 DeepMind CEO predicts AGI within three years
  • 🤖 Endava uses Codex to build Agent-driven org, reducing requirement analysis from weeks to hours
  • 🤖 Google DeepMind launches Gemini Embedding 2, multimodal embedding model for AI search
  • 🤖 AI startup OpenClaw founder and OpenAI product lead to speak together at VivaTech
  • 🤖 AWS+LangSmith evaluates Deep AI Agents: 5 modes & million-token models
  • 🤖 LLMs still believe false statements after warnings
Unlock all 11 headlines + deep analysis →Free 3-day trial · cancel anytime
Browse all past briefings →