Friday, September 25, 2026

Several AI coding agents experienced detected pricing changes recently. GitHub Copilot's typical pricing increased from $10 to $15/month, with the floor also shifting to $15/month GitHub Copilot pricing change detected: typical $10 → $15/mo, floor $10 → $15/mo. OpenAI Codex saw its floor price move from $0 to $8/month OpenAI Codex pricing change detected: floor $0 → $8/mo, while Cursor's floor pricing similarly moved from $0 to $20/month Cursor pricing change detected: floor $0 → $20/mo. In contrast, Claude Code's typical pricing decreased from $20 to $17/month Claude Code pricing change detected: typical $20 → $17/mo. These detected changes are pending human verification.

In agent updates, OpenAI Codex released version 0.157.0, which included GPT-6 Sol and Luna models and introduced Amazon Bedrock support 0.157.0. GitHub Copilot implemented a new global default policy for generally available features and supported client capabilities in enterprise and organization settings Default Enablement of Copilot Features for Copilot Business and Enterprise. Additionally, Gemini introduced Gemini 3.8 Live with Live Avatar, bringing near real-time visual presence to its conversational AI Introducing Gemini 3.8 Live with Live Avatar.

The community also saw various projects and discussions around enhancing AI coding agent workflows. Tools like the "Vibe Coding Production Kit" aim to provide a production workflow for AI coding agents Vibe Coding Production Kit – a production workflow for AI coding agents. Several new interfaces and verification systems were presented, including "Radix," a visual UI for programming agents that supports Claude Show HN: Radix – Visual UI for agentic programming, and "Whiteboard," an open-source IDE for collaborative software design between humans and agents like Claude Code and Codex Show HN: Whiteboard (YC W26) – An open-source IDE for thoughtful software design. For code review, "Critic" allows reviewing code with the agent that wrote it, supporting Codex or Claude Code Show HN: Critic – Review code with the agent that wrote it, while "Canary" offers independent verification for AI code from Claude and Codex using agent swarms Show HN: Canary (YC) – Independent verification for AI code. GitHub also discussed scenarios where a chat interface might not be the optimal UI for developers, suggesting alternatives like canvases When chat is the wrong UI.