Head-to-head · Oct 15, 2026
Claude Code vs Codex vs Cursor
Scored for the job: Best AI coding agents
⚠ Preliminary, based on public research, hands-on test pending. These scores come from a desk review of public sources (official pricing pages, docs, changelogs), linked on each tool page. Rex hasn't run the hands-on tests yet, so tiers shown are projected. Under our rubric, a tool's official tier is capped at B until it's tested.
| Criterion | Claude Code | OpenAI Codex | Cursor |
|---|---|---|---|
| Projected tier | S86/100 | A84/100 | A80/100 |
| Output quality on the job test /35 | 31 | 31 | 29 |
| Speed & friction /15 | 13 | 13 | 14 |
| Value / price /15 | 11 | 10 | 9 |
| Control & ownership /10 | 8 | 7 | 7 |
| Momentum & community /10 | 10 | 10 | 9 |
| Reliability & trust /10 | 8 | 8 | 7 |
| Builder-fit /5 | 5 | 5 | 5 |
| Free tier | No | Yes | Yes |
| Entry paid plan | Pro: $20/mo ($17/mo yearly) | Plus: $20/mo | Plus: $8/mo |
| Verdict | The terminal-native agent many builders reach for first; maxed builder-fit (CLI, MCP, hooks). Heavy use pushes you to the $100-200 Max plans. | Buffed in Sep 2026 with GPT-6 Astra, and included in every ChatGPT plan. Plus-tier Astra limits are tight. | Slickest editor UX. Pricing reshuffled on Oct 12, 2026 into The X Subscription; third-party models now billed at API list price. |
Pricing checked Oct 15, 2026. Sources on each tool page.
Quick take (preliminary): Claude Code and Codex are neck and neck at the top of our desk review. Both are bundled into $20 chat subscriptions and scale to $100-200 plans for heavy use. Cursor keeps the best editor UX, but its Oct 12, 2026 move to The X Subscription means third-party models (like Claude or GPT) are now billed separately at API list price on new plans.
- Terminal-first: Claude Code, Codex CLI.
- Editor-first: Cursor.
- Free option: Codex (ChatGPT Free, limited) and Cursor Free (limited). Claude Code needs Pro.
The head-to-head test on a shared repo task pack is queued for week 2.