The Battle of AI Coding Tools: Claude Code vs. Codex CLI
In the world of AI coding tools, two prominent players have emerged: Anthropic’s Claude Code and OpenAI’s Codex CLI. However, it seems that Codex CLI is winning over developers more than Claude Code, due in part to Anthropic’s strict usage policies and actions against developers.
Both Claude Code and Codex CLI offer similar functionalities, allowing developers to leverage AI models in the cloud for coding tasks. Released around the same time, these tools aim to capture the attention of developers and establish themselves in the market.
One key difference between the two tools is their licensing. Codex CLI is available under an Apache 2.0 license, allowing for distribution and commercial use. On the other hand, Claude Code is tied to Anthropic’s commercial license, restricting modifications without explicit permission.
Anthropic further complicated matters by obfuscating the source code for Claude Code, making it difficult for developers to access and modify. When a developer managed to de-obfuscate the code and shared it on GitHub, Anthropic responded with a DMCA complaint to have the code removed.
The developer community expressed disappointment with Anthropic’s actions, contrasting them with OpenAI’s more developer-friendly approach with Codex CLI. OpenAI has been quick to incorporate developer suggestions into Codex CLI, including the ability to use AI models from other providers, such as Anthropic.
Despite the backlash, Anthropic has not provided any comments on the issue. It’s worth noting that Claude Code is still in beta and may see changes in the future, including a more permissive licensing agreement. Companies often have valid reasons for obfuscating code, such as security concerns.
This turn of events marks a positive moment for OpenAI, which has previously focused on proprietary products over open-source releases. It suggests a shift in the company’s strategy, as CEO Sam Altman acknowledged the importance of open source in the industry.
