Scoutari

AI builder 的決策情報台。

瀏覽

  • 今日情報
  • Skill 榜單
  • AI 週報
  • 關於與編輯方針
  • RSS

主題

  • 模型發布
  • API 與價格
  • MCP 與 Skills
  • AI Coding
  • 中國模型
  • 開源模型
  • Agent
  • Infra / 成本
  • 值得試用
© 2026 Scoutari · scoutari.com摘要由 AI 輔助生成,並始終附上原文連結。
情報收藏週報設定
SScoutariAI Builder Intel · decision desk
Skill RadarSavedPersonalizeSubmit sourceLog inSign up
Sign up
Skill RadarSavedPersonalizeSubmit sourceLog in

Model releases

Releases and upgrades across Claude, GPT, Gemini, Qwen, DeepSeek, Kimi, and more.

API & pricingMCP & SkillsAI codingChinese modelsOpen sourceAgentsInfra & costMultimodal & imageAI videoVoice AIAI hardwareRobotics & embodiedAI safetyResearch & papersEnterprise AIAI fundingPolicy & regulationWorth trying
  • Mon, July 6Open Source

    MaxKB: Open-Source Enterprise Agent Platform Trending on GitHub

    MaxKB is an open-source enterprise agent platform gaining significant traction on GitHub.

  • Mon, July 6Tools

    planning-with-files: Persistent File-Based Planning for AI Coding Agents

    OthmanAdi/planning-with-files is a file-based persistent planning tool designed for AI coding agents and long-running agent tasks.

  • Mon, July 6Tools

    Claude Code Skill Humanizer: Open-Source Tool to Remove AI-Writing Traces Tops GitHub Trending

    A Claude Code skill called blader/humanizer reached 27,579 stars on GitHub, with 826 new stars in 7 days, topping the trending chart.

  • Mon, July 6Tools

    garrytan/gstack Explodes in Popularity: 23 Tools Simulate Team Roles from CEO to QA

    GitHub trending repo garrytan/gstack offers Garry Tan's Claude Code config with 23 tools covering six team roles from CEO to QA.

  • Mon, July 6Tools

    PM Skills Marketplace Launches with 100+ Agent Skills

    GitHub trending project pm-skills launches PM Skills Marketplace with over 100 agent skills, commands, and plugins.

  • Mon, July 6Open Source

    claude-mem: Persistent context across sessions for all agents

    claude-mem captures all agent actions in a session, compresses them with AI, and injects relevant context into future sessions.

  • Mon, July 6Tools

    earendil-works/pi: A unified LLM API and Agent loop toolkit

    GitHub trending repo earendil-works/pi offers an AI toolkit with unified LLM API, Agent loop, TUI, and coding agent CLI, gaining 1,423 stars in 7 days (total 67,794).

  • Mon, July 6Tools

    Understand-Anything: Turn Any Code into Interactive Knowledge Graph

    Egonex-AI/Understand-Anything converts any codebase into an explorable, searchable, and queryable interactive knowledge graph, earning over 1,515 GitHub stars in a week.

  • Mon, July 6Tools

    alirezarezvani/claude-skills releases 337 skills and plugins supporting 11 code agents

    GitHub trending project alirezarezvani/claude-skills offers 337 Claude Code skills, agent skills, and plugins covering engineering, marketing, and more, supporting 11 code agents including Claude Code, Codex, Gemini CLI, and Cursor.

  • Mon, July 6Tools

    Claude Code Hits GitHub Trending, Terminal Agent Coding Tool Gets 136K Stars

    Anthropic's Claude Code, a terminal-based agent coding tool, reaches GitHub trending with 136K stars.

  • Mon, July 6Open Source

    ECC Launches Agent Performance Optimization Framework for Claude Code, Codex, Cursor, etc.

    GitHub trending project ECC launches as an agent performance optimization system, offering skills, instincts, memory, and security.

  • Mon, July 6Research

    GitHub trending repo system_prompts_leaks aggregates AI model system prompts

    GitHub repo system_prompts_leaks collects and updates system prompts from Anthropic, OpenAI, Google, and other AI models.

  • Sun, July 5Tools

    LlamaIndex Launches legal-kb Demo: Agent-Based Index v2 Document Knowledge Base

    LlamaIndex releases legal-kb, a reference app enabling AI agents to access Index v2 document knowledge bases via tools like retrieve, find, read, and grep, resembling a file system.

  • Sun, July 5Research

    2026 Guide to Open-Source PDF Structured Extraction Models

    This article explains how to use open-source models to convert enterprise data from PDFs, scans, and slides into structured JSON for LLMs and agents.

  • Sun, July 5Tools

    Alibaba Reportedly Bans Claude Code, Labels It High-Risk Software

    Alibaba has reportedly banned employees from using Claude Code, classifying it as high-risk software.

  • Sun, July 5Tools

    Alibaba Reportedly Bans Employees from Using Claude Code

    Alibaba reportedly prohibits employees from using Anthropic's Claude Code coding tool.

  • Sat, July 4Model/API

    Mistral AI: The OpenAI Competitor You Need to Know

    Mistral AI offers partially open-source AI models and has raised substantial funding since its founding in 2023, aiming to put frontier AI in everyone's hands.

  • Sat, July 4Research

    Anthropic Dev Shares Fable 5 Prompting Tips: Find Blind Spots First

    Anthropic developer Thariq Shihipar shares prompting strategies for Fable 5, urging developers to systematically identify their knowledge blind spots before delegating tasks to Claude.

  • Sat, July 4Tools

    Claude Code v2.1.201 removes mid-conversation system role harness reminders

    Claude Code v2.1.201 drops mid-conversation system role harness reminders for Claude Sonnet 5 sessions.

  • Sat, July 4Tools

    Claude Code v2.1.200 Fixes Key Background Agent and Permission Mode Issues

    Claude Code v2.1.200 fixes multiple issues with background agents, permission modes, and screen readers, and improves the installation script.

  • Fri, July 3Model/API

    DeepSeek Peak Pricing, Goldman Sachs Says China's LLM Competition Turns Rational

    DeepSeek raises API prices during peak hours, signaling a shift from subsidy-driven competition to sustainable pricing in China's AI model market.

  • Fri, July 3Research

    Using DSPy to Evaluate and Improve Datasette Agent SQL System Prompt

    A study uses DSPy to evaluate and improve Datasette Agent's SQL system prompt.

  • Fri, July 3Model/API

    Anthropic cuts Claude Code system prompts by 80%

    Anthropic reduced Claude Code's system prompt length by 80% because the Fable 5 model prefers smaller prompts.

  • Thu, July 2Tools

    Safari New MCP Server Lets Coding Agents Inspect and Debug Websites

    Safari launches a new MCP server enabling coding agents to inspect and debug websites.

  • Thu, July 2Tools

    Claude Code v2.1.198: Chrome GA, Agent Notifications, AWS Upstream, Fixes

    Claude Code v2.1.198 makes Chrome's Claude generally available, adds background Agent notifications, /dataviz skill, AWS upstream support, and fixes multiple issues.

  • Thu, July 2Model/API

    Anthropic redeploys Claude Fable 5 on July 1 after US export controls ease, adds security classifier

    Anthropic redeployed Claude Fable 5 on July 1 after US export controls were lifted, and added a safety classifier.

  • Thu, July 2Tools

    Google June 2026 AI Update Announcement

    Google announced a series of latest AI news in June 2026.

  • Wed, July 1Tools

    Google Agent Assistant Gemini Spark Now Supports Mac

    Google's 24/7 agent assistant Gemini Spark is now available on Mac.

  • Wed, July 1Tools

    Google Launches New Smart Speaker, But Gemini Not Ready

    Google releases a new smart speaker, but its Gemini AI is not ready to support the device.

  • Wed, July 1Model/API

    NVIDIA Releases Nemotron-Labs-TwoTower Diffusion Language Model

    NVIDIA released an open-weight diffusion language model built on a frozen autoregressive backbone.

  • Wed, July 1Model/API

    Google AI Launches TabFM: Hybrid Attention Table Foundation Model for Zero-Shot Classification and Regression

    Google Research released TabFM, a foundation model for tabular data that enables zero-shot classification and regression via in-context learning.

  • Wed, July 1Tools

    Ollama v0.31.1 Boosts Gemma 4 Token Generation by Nearly 90% on Apple Silicon with Multi-Token Prediction

    Ollama v0.31.1 uses multi-token prediction to significantly accelerate Gemma 4 inference on Apple Silicon.

  • Wed, July 1Model/API

    Anthropic's Long-Delayed Claude Fable 5 Gets Greenlight

    Anthropic announced that Claude Fable 5 will relaunch after weeks of negotiations with the Trump administration.

  • Wed, July 1Model/API

    Anthropic: US Commerce Dept Lifts Export Controls on Claude Fable 5 and Mythos 5

    Anthropic received notice that the US Commerce Department lifted export controls on Claude Fable 5 and Mythos 5, restoring access tomorrow.

  • Wed, July 1Tools

    Anthropic Launches Claude Science Flagship

    Anthropic announced Claude Science, a new flagship product supporting scientific research.

  • Wed, July 1Model/API

    Claude Sonnet 5 Released: New Tokenizer Boosts Efficiency but Costs 30% More

    Anthropic releases Claude Sonnet 5 with performance near Opus 4.8 at lower price, but new tokenizer raises effective cost by ~30%.

  • Wed, July 1Tools

    Cline v4.0.5 Adds Claude Sonnet 5 Support Across Providers

    Cline v4.0.5 adds Claude Sonnet 5 support, covering Anthropic, Bedrock, and other providers.

  • Wed, July 1Tools

    Claude Code Default Model Upgraded to Claude Sonnet 5, Natively Supports 1M Token Context

    Claude Code v2.1.197 introduces Claude Sonnet 5 as the default model with native 1M token context window and promotional pricing until August 31.

  • Tue, June 30Research

    OpenAI Launches GeneBench-Pro Benchmark for AI in Genomics and Science

    OpenAI releases GeneBench-Pro, a new benchmark using complex real-world datasets to test AI performance in genomics, biology, and scientific research.

  • Tue, June 30Open Source

    DeepReinforce Launches Ornith-1.0: Self-Constructed LLM for Agentic Coding

    DeepReinforce releases its first open-source model, Ornith-1.0, based on Gemma 4 and Qwen 3.5, achieving state-of-the-art results across coding benchmarks.