Cursor, Claude Code, and Codex are merging into one AI coding stack nobody planned
As the frontier models converge, the coding-tool contest has moved to the agent wrapper — and developers are combining tools rather than choosing one. The Cursor-plus-Claude-Code pairing reports the highest self-rated productivity; OpenAI now ships a plugin that runs inside Anthropic's Claude Code. Rival tools are quietly becoming one interoperating stack.
Convergence at the model layer pushed differentiation up to the harness. With frontier models scoring similarly, the agent wrapper — the terminal, the editor integration, the orchestration — now decides the experience. Claude Code owns the programmable terminal, Cursor the in-editor speed, Codex and Devin the run-it-without-watching autonomy, and each occupies a distinct role rather than fighting for one seat.
Interoperation is the surprise. OpenAI publishing an official plugin that runs inside Anthropic's Claude Code is competitors treating each other's surfaces as platforms — a sign the market has decided developers will mix tools, so the smart move is to be present in the others' workflows rather than wall them off. The stack nobody planned is emerging from that logic.
The productivity data validates the stacking. The Cursor-plus-Claude-Code combination reporting the highest self-rated productivity — Cursor for inline edits, Claude Code for shipping complete features — shows the tools are genuinely complementary. The winning configuration is a stack, which is why the vendors are building for coexistence, not conquest.
The New Stack — Cursor, Claude Code, and Codex are merging into one AI coding stack nobody planned → · Morph — Best AI coding agents (August 2026): scored leaderboard →