Best AI Coding Tools 2023

Claude Code is praised for its ability to understand and refactor codebases and handle multi-file reasoning, despite sometimes rewriting more than requested. "Claude Code - Really reads the codebase. Refracting, catching bugs in replies that I didn't write, migrations."

OpenCode is frequently mentioned as a strong CLI tool for coding, with some Users citing it as their absolute leader. "Opencode is the absolute leader atm.... using it from 1 year less or more and is the #1 cli tool."

Codex (often used with ChatGPT business/plus) is considered excellent for precision fixes and general coding assistance, especially when integrated into environments like VS Code. "I'm using Github Copilot (pro+) and Codex in VS Code and i'm very happy! I use them everyday."

Diverse Tool Use Many Users advocate for using a combination of tools, switching between them based on the specific task, such as using Claude for multi-file reasoning and Cursor for in-editor changes. "I switch between different tools depending on the task, that's an important rule for being productive."

Agent Mode For more complex tasks, Users utilize agent mode workflows where AI tools plan and execute changes, often requiring careful review and refinement. "Agent mode for the actual implementation, but I still make it walk me through the plan before touching anything, otherwise it goes off and refactors three files you didn't ask about."

Context Management Successfully using AI tools involves carefully providing context and setting constraints to ensure relevant and accurate outputs. "The part that changed the most for me isn't the model, it's how much I front-load context."

Cost vs. Quality While Claude Code offers high quality, some Users find its usage limits restrictive, leading them to explore more cost-effective alternatives like GLM 5.1 or combinations of tools. "Claude Code felt amazing quality-wise, but the limits disappeared faster than gas in an old Opala with a 6-cylinder engine."

Open-source and Local Models Some Users experiment with local LLMs and open-source options like OpenCode to manage costs and gain more control over their coding environment. "I got OpenCode running with my local LLM and Chrome dev tools."