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

API & pricing

Model API capability changes, pricing moves, token and rate policies.

Model releasesMCP & SkillsAI codingChinese modelsOpen sourceAgentsInfra & costMultimodal & imageAI videoVoice AIAI hardwareRobotics & embodiedAI safetyResearch & papersEnterprise AIAI fundingPolicy & regulationWorth trying
  • 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

    n8n workflow automation platform rises 865 stars on GitHub with native AI capabilities

    n8n is a fair-code workflow automation platform with native AI capabilities, supporting visual building and custom code, self-hosted or cloud, with over 400 integrations, recently gaining 865 stars on GitHub Trending.

  • 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 6Agent

    mvanhorn/last30days-skill: Cross-platform AI Agent skill aggregating Reddit, X, YouTube sources for summaries

    A GitHub trending project that researches any topic across multiple platforms and synthesizes evidence-based summaries.

  • Mon, July 6Tools

    CodeGraph: Pre-indexed Code Knowledge Graph, Auto-Syncs Changes, Supports Multiple AI Agents

    CodeGraph is a pre-indexed code knowledge graph that auto-syncs on code changes and supports AI agents like Claude Code, Codex, and Gemini.

  • Mon, July 6Tools

    Open Design: Local-First Desktop App Turns Code Agents into Design Engines

    Open Design is an open-source, local-first desktop alternative to Claude Design, enabling code agents to generate designs and export to HTML/PDF/PPTX/MP4.

  • Mon, July 6Open Source

    AgentSkills Specification & Docs Open Source Project Tops GitHub Trending

    The GitHub trending project agentskills/agentskills provides specifications and documentation for Agent Skills, gaining 2262 stars in 7 days.

  • Mon, July 6Tools

    Headroom: Open-Source Tool to Compress Tool Outputs, Logs, Files, and RAG Chunks by 60-95% Tokens Without Quality Loss

    Headroom is an open-source project that compresses tool outputs, logs, files, and RAG chunks before sending them to LLMs, reducing token usage by 60-95% while maintaining answer quality.

  • Mon, July 6Tools

    DeusData/codebase-memory-mcp: High-Performance Code Intelligence MCP Server Indexes Codebase as Persistent Knowledge Graph

    High-performance MCP server indexing codebases into persistent knowledge graphs with millisecond processing for average repos.

  • Mon, July 6Tools

    JuliusBrussee/caveman: Caveman mode cuts token usage by 65%

    JuliusBrussee/caveman is a Claude Code skill that uses caveman-style speech to reduce token consumption by 65%.

  • Sun, July 5Model/API

    Mistral CEO Warns: Proprietary AI Models Let Labs Peek at Your Business Processes

    Mistral co-founder Arthur Mensch warns enterprises against relying on closed AI models, stating that AI labs are storing more customer data and may even use it to compete against customers.

  • Sun, July 5Research

    From Prompt Engineering to Tokenmaxxing: AI Productivity Trends

    Explores AI productivity trends from prompt engineering to tokenmaxxing.

  • 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 Best Practice: Let Fable Decide When to Use Low-Cost Models

    Simon Willison, following Claude Code team advice, lets Fable autonomously decide to delegate coding tasks to low-cost subagents to save tokens.

  • 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 3Model/API

    Kuaishou Stock Rises on Tencent's $2.8B Kling AI Investment

    Kuaishou's stock surged after Tencent joined a $2.8 billion funding round for its Kling AI subsidiary.

  • Fri, July 3Agent

    Claude Code v2.1.199 Fixes Sub-Agent & Daemon Stability

    Claude Code v2.1.199 fixes sub-agent error handling, background daemon crashes, and SSL error prompts, improving stability.

  • Fri, July 3Tools

    llm-coding-agent 0.1a0: CLI & Python API Coding Agent

    Simon Willison releases llm-coding-agent 0.1a0, a simple coding agent based on the LLM library.

  • Fri, July 3Tools

    Weber July 4th Sale: Grills & Griddles at All-Time Lows

    Weber offers massive discounts on grills, smokers, griddles, and accessories before July 4th, with prices hitting record lows.

  • Thu, July 2Tools

    Scritty: Shared Searchable Memory for Every AI Coding Agent

    Scritty provides shared, searchable memory for each AI coding agent.

  • Thu, July 2Agent

    Build a Serverless A2A Gateway on AWS for Agent Discovery, Routing, and Access Control

    A blog on AWS Machine Learning Blog demonstrates how to build a serverless A2A gateway on AWS, hosting multiple agents under a single domain via path-based routing /agents/{agentId}, with standard A2A clients working without modification.

  • Wed, July 1Model/API

    Claude Sonnet 5 Hidden Price Hike: Per-Task Token Consumption Up 40%, Real Cost Doubles

    Claude Sonnet 5 surpasses Opus 4.8 on some tasks, but its per-task token consumption increased by ~40%, nearly doubling actual costs—continuing Anthropic's pattern of hidden price increases.

  • 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

    Anthropic Claude Sonnet 5 vs Sonnet 4.6 vs Opus 4.8: Agentic Coding Benchmarks, API Pricing & Value

    Anthropic's Claude Sonnet 5 narrows the gap with Opus 4.8 on agentic coding tasks while maintaining lower Sonnet-tier pricing.

  • 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 1Model/API

    Anthropic Launches Cheaper Claude Sonnet 5 for Agents

    Anthropic introduces Claude Sonnet 5 as a cost-effective model for running AI agents.

  • Wed, July 1Model/API

    Anthropic Launches Claude Sonnet 5 as Cheaper Agent Option

    Anthropic releases Claude Sonnet 5 with stronger agentic capabilities, lower price, and improved safety, positioning it as a cheaper alternative to Opus, GPT-5.5, and Gemini Pro.

  • 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 30Model/API

    Vibe coding platform Base44 launches own AI models to boost startups' defenses

    Wix's vibe coding platform Base44 begins releasing its own AI models, aiming to eventually surpass frontier models.

  • Tue, June 30Tools

    NVIDIA BioNeMo Agent Toolkit Transforms Biomolecular Models into AI Agent Skills

    NVIDIA open-sources BioNeMo Agent Toolkit to convert biomolecular models into callable skills for AI agents, boosting task completion from 57.1% to 100% in tests.

  • 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.

  • Sat, June 27Open Source

    Meta Releases Astryx: Open-Source React Design System with CLI and MCP Server for AI Agents

    Meta released Astryx, an open-source React design system built on StyleX, integrating CLI and MCP server for both human engineers and AI agents.

  • Sat, June 27Tools

    Building SFT Data from NVIDIA Open-SWE-Traces: Trace Parsing, Patch Analysis, Token Budget, Tool Usage Metrics

    A tutorial on streaming the NVIDIA Open-SWE-Traces dataset from Hugging Face in Google Colab to efficiently process agentic software engineering traces and generate a subset for fine-tuning.

  • Sat, June 27Tools

    MRAgent Memory System Cuts Token Usage 27x

    MRAgent reduces LLM token consumption by up to 27x via optimized agent memory management.

  • Sat, June 27Model/API

    OpenAI Launches GPT‑5.6 Series: Sol, Terra, Luna in Limited Preview

    OpenAI began limited preview of GPT‑5.6 series (Sol, Terra, Luna) with full rollout planned within weeks.

  • Fri, June 26Model/API

    OpenAI Internal Codex Output Surges: Research Up 56x

    Median Codex output tokens at OpenAI have surged 56x in research since Nov 2025.

  • Fri, June 26Tools

    GitHub Copilot Agentic Harness: Cross-Model Performance and Efficiency

    GitHub evaluates the Copilot agentic framework across benchmarks, showing token efficiency and multi-model flexibility.

  • Fri, June 26Research

    Microsoft Announces August Xbox Price Hike, Prime Day Discounts Remain

    Microsoft will raise prices on all Xbox models in August, the second increase in under a year.

  • Fri, June 26Agent

    Retrofit, Not Rebuild: Agentic Overlays for Enterprise Services

    AWS proposes agentic overlays to transform legacy REST services into A2A-capable agents with MCP compatibility.