¿El mejor CLI en 2025? Qwen, Gemini o Claude

¿El mejor CLI en 2025? Qwen, Gemini o Claude

by Codemancers - Inteligencia Artificial

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Codemancers - Inteligencia Artificial 👥 2K 📅 October 7, 2025 ⏱ 66 min 👁 394 📄 expert opinion 🧭 2026-08-15
Available in: English (current) Français

Keywords

AI codingCLIQwenGeminiClaude

Summary

In this episode of Codemancers, the hosts compare three AI coding assistants running from the terminal: Qwen Code (cloud version), Gemini CLI, and Claude Code. They set up a practical task: building a website for their podcast, including a feature to display the latest six YouTube videos. They use the same prompt for all three tools, with rules to limit internet access and comments. The comparison focuses on speed, code quality, and ability to handle a broad task. Qwen Code finishes first but produces a basic page without styling. Claude Code produces a more complete and functional site, while Gemini CLI struggles and fails to deliver a working result. The hosts also discuss their personal experiences: one prefers Claude for coding and Gemini for analyzing unfamiliar code, while the other notes improvements in ChatGPT 5. They conclude that Claude Code is the winner for this benchmark, but emphasize that the prompt was intentionally vague and that better prompts could improve results. They also mention that Qwen Code local was excluded due to slowness. The episode ends with a teaser for a future segment where one host will test the tools on bug-fixing in existing code.

195 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides practical, hands-on insights into the performance of three AI coding assistants. The hosts demonstrate real usage, showing the tools’ outputs and discussing their strengths and weaknesses. The argumentation is based on personal experience and a single benchmark, which limits generalizability. They acknowledge the limitations of their methodology, noting that the prompt was intentionally vague and that better prompting could yield different results. The discussion is balanced, with each tool given a fair chance, and the hosts share their individual preferences and workflows. However, the lack of rigorous testing and the small sample size reduce the scientific value of the conclusions.

Scientific Rigor, Source Quality, Title Accuracy

The video does not cite external sources or references. The only links provided in the description are to the podcast’s distribution platforms (Spotify, Apple Podcasts, Ivoox) and their website, which are not directly related to the content. The title accurately reflects the content, as the video indeed compares Qwen, Gemini, and Claude CLI tools. The hosts demonstrate a good understanding of the tools and provide practical advice, but the lack of citations and the informal methodology limit the scientific rigor. No comments were provided for analysis.

204 words

Title / Content Match

The title accurately reflects the content: a comparison of three CLI coding assistants in 2025.

Quality & Reliability

6/10

The video presents a practical comparison of three AI coding assistants (Qwen Code, Gemini CLI, Claude Code) based on personal experience and a single benchmark task. The methodology is informal and lacks rigorous controls, but the hosts provide detailed observations and acknowledge limitations. The information is useful for practitioners but not scientifically validated.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Concurring Sources

  • Claude Code documentation — Official documentation for Claude Code, which aligns with the video's positive assessment of Claude Code's capabilities.
  • Gemini CLI GitHub repository — Official repository for Gemini CLI, providing information that supports the video's description of its features.
  • Qwen Code documentation — Official documentation for Qwen models, which aligns with the video's mention of Qwen Code's performance.

Dissenting Sources

  • No discordant sources found — The video does not cite any sources that contradict its claims.

Contribution & Novelties

The video offers a practical, real-world comparison of three AI coding assistants, providing insights into their performance on a typical web development task. It highlights the strengths and weaknesses of each tool, such as Claude Code’s superior code generation and Gemini’s ability to handle large contexts but tendency to get lost. The hosts also share their personal workflows, suggesting a division of labor: using Gemini for code analysis and Claude for coding. This practical advice is valuable for developers considering these tools.

Pour aller plus loin :

131 words

Radar Profile

The radar profile shows moderate scores across all dimensions, with quantity of information and technical level being relatively higher, while reliability is lower due to the informal methodology. This suggests the video is informative and technically detailed but lacks scientific rigor.

Reliability 5/10