Companies test Codex, but Claude Code remains engineers’ preferred choice
Key Takeaways
- •Claude Code was responsible for roughly 4% of public GitHub commits by February 2026 and reached a peak of 326,000 commits per day in mid-March 2026.
- •Claude Code scored 72.7% on verified benchmarks, ahead of Codex at 69.1%.
- •Developers reported that Claude Code performs especially well in long, multi-step coding sessions, UI work, and plugin-supported tasks.
- •Codex has an efficiency advantage for contained delegation-style work because it can be cheaper per task.
- •Enterprises in financial services, healthcare technology, and other regulated industries are increasingly adopting Claude Code, often alongside Codex in hybrid workflows.
Across enterprise engineering teams evaluating the latest generation of AI coding agents, Anthropic’s Claude Code has emerged as the tool developers actually want to use. OpenAI’s Codex is receiving plenty of trials, but when engineers are given a choice, Claude Code continues to win support, underscoring how adoption is being shaped not just by model quality but by day-to-day workflow fit and task complexity.
The numbers tell the story
By February 2026, Claude Code was responsible for roughly 4% of all public GitHub commits, averaging about 135,000 commits per day. That figure climbed to a daily peak of 326,000 commits by mid-March 2026.
On verified benchmarks, Claude Code scored 72.7%, compared with 69.1% for Codex.
Developer reports throughout 2026 have repeatedly highlighted Claude Code’s strength in long-session contexts, the kind of extended, multi-step coding work in which an AI agent must remember what it did 40 steps earlier without losing the thread. It also receives strong marks for UI-related tasks and for its plugin ecosystem.
Codex, however, has an advantage in one area: price-per-task efficiency. For delegation-style work that is well scoped and relatively contained, Codex can be the more economical option.
Enterprise adoption is shifting
Anthropic reportedly surpassed OpenAI on enterprise adoption metrics beginning in early 2025, particularly in developer satisfaction and agentic task completion rates.
Claude Code is gaining traction in enterprise contracts, especially in environments where consistent output quality and compliance are important. Financial services, healthcare technology, and regulated industries are increasingly adopting Anthropic’s offering, where teams often weigh reliability and governance alongside raw coding speed.
Several firms are also quietly adopting hybrid workflows that use both tools depending on the task. Claude Code is assigned the more complex, context-heavy work, while Codex handles routine, cost-sensitive tasks. That split reflects a broader enterprise pattern: teams are testing multiple agents, then matching each one to the kind of work it handles best rather than standardizing on a single assistant.