Agentic Coding Weekly

Welcome to the Agentic Coding Weekly Newsletter! Get focused updates on agentic coding tools, models, and workflows. No hype, no gossip. Just what helps you ship. Weekly issues every Monday. Subscribe ⬇️ agenticcodingweekly.com

Issue 18 is out, covering agentic coding updates from last week: OpenAI releases GPT-5.4 Mini and Nano, Cursor launches Composer 2, a practical multi-agent workflow and three reads on specs, codegen, and real-world usage Read: agenticcodingweekly.com/p/acw-18-gpt-5…

If you're still typing every prompt to coding agents, you're making things hard for yourself Use one of these open-source voice typing tools with Claude Code, Codex, Cursor, Copilot, or any coding agent to give richer context with less effort Read: agenticcodingweekly.com/p/5-best-open-…

Issue 17 is out, covering agentic coding updates from last week: 1M context for Claude, 2x usage promo, JetBrains Air, /btw command, voice mode, and 8 levels of agentic engineering Read: agenticcodingweekly.com/p/acw-17-1m-co…

Claude Code now has built-in voice mode for dictation So I compared it with what I typically use, gpt-4o-transcription model through VoiceInk gpt-4o-transcription was noticeably more accurate and Claude Code voice mode struggled more, especially with technical terms

2x usage limits are active for Claude during off-peak hours from March 13 - 27 Promotion applies automatically across Claude Code, Cowork, and the web/mobile applications Bonus usage are excluded from standard weekly limits

2x Claude usage for the next 2 weeks. Did the timezone math for you.

1 million context window: Now generally available for Claude Opus 4.6 and Claude Sonnet 4.6.

1M token context window is now GA for Claude Opus 4.6 and Sonnet 4.6 no more long-context pricing premium on Max, Team, and Enterprise plans, Opus 4.6 sessions now automatically utilize the 1M context

Issue 16 is out, covering agentic coding updates from last week: GPT-5.4's 1M context, Claude Code scheduled tasks and voice mode, Gemini 3.1 Flash-Lite, agentic engineering patterns, and the debate over AI-assisted relicensing

GPT-5.4 has 1M token context window and a new tool search feature that cuts token usage nearly in half for tool-heavy workflows Matches GPT-5.3-Codex on SWE-Bench Pro (57.7%) while running at lower latency, and the /fast mode in Codex pushes token velocity up to 1.5x

GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT. GPT-5.4 is also now available in the API and Codex. GPT-5.4 brings our advances in reasoning, coding, and agentic workflows into one frontier model.

GPT-5.4 is here. Native computer-use capabilities. Up to 1M tokens of context in Codex and the API. Best-in-class agentic coding for complex tasks. Scalable tool search across larger ecosystems. More efficient reasoning for long, tool-heavy workflows. openai.com/index/introduc…

Issue 15 is out, covering agentic coding updates from last week: Claude Code remote control, Cursor cloud VMs, RTK token-saving proxy, /simplify and /batch skills, Codex voice dictation, and Cloudflare rebuilt Next.js from scratch Read: agenticcodingweekly.com/p/acw-15-claud…

Check answers and explanations here ⬇️ agenticcodingweekly.com/p/agentic-codi…

Claude Code ships with built-in ones like Explore and Plan, but you can also create custom subagents tailored to your workflow Test your knowledge of Claude Code's subagent system and custom subagent configuration with these 5 questions

Issue 14 is out, covering agentic coding updates from last week: Claude Sonnet 4.6 and Gemini 3.1 Pro releases, Claude Code worktree support, and spec-driven development from Shisa AI's CTO Read: agenticcodingweekly.com/p/acw-14-claud…