Waper
Use on the web

AI coding tools

What changed in Claude Code, Codex, Cursor, Copilot and other coding agents, and what actually works when building with them.

  • Official
  • Public
  • 0 followers
Sign up to followCustomize it
You're reading the brief from Tue, Oct 27.
Back to the latest brief

Cognition introduced FrontierCode 2.0, a coding-agent benchmark focused on whether open-source maintainers consider generated changes mergeable.

AI coding tools1 item

Ask Kiki about this brief

Sign in to ask questions about this brief.

Recent briefs

  1. Tue, Nov 32 itemsA quiet edition focuses on practical applications: Codex in physical engineering workflows and Copilot in maintaining a solo developer’s Minecraft project.
  2. Mon, Nov 22 itemsDevin added delegated budget controls for enterprise AI usage, while Qodo described bringing company context and coding standards into Codex-based code review.
  3. Sun, Nov 1Nothing worth reporting
  4. Sat, Oct 31Nothing worth reporting
  5. Fri, Oct 303 itemsCodex 0.166.0 adds GPT-6.1 Terra to its model catalogs and improves keyboard-based transcript navigation. OpenAI also introduced API-credit benefits for eligible ChatGPT subscribers, potentially changing the cost of API-based development workflows.
  6. Thu, Oct 298 itemsGitHub’s biggest changes were automatic model routing, Copilot access bundled with eligible Enterprise arrangements, and generally available Agentic Workflows. Elsewhere, Claude Code expanded monitoring support, while practical reports explored offline agents and agent-built games.
  7. Wed, Oct 285 itemsGitHub made Copilot Fusion and Agentic Workflows generally available and expanded Copilot access for eligible Enterprise customers. OpenAI also announced a ServiceNow partnership expansion that brings ServiceNow workflows into Codex.

Who and what it tracks

  • CodexProduct7
  • Claude CodeProduct6
  • GitHub CopilotProduct6
  • DevinProduct4
  • Gemini CLIProduct2