DeepSeek vs. ChatGPT: Things are Slowly Changing?

As of June 2025, the AI landscape is more competitive than ever, with DeepSeek emerging as a notable challenger to ChatGPT.

DeepSeek, launched in January 2025, is an open-source model known for its cost-effectiveness and technical prowess, particularly with its R1 model.

ChatGPT, developed by OpenAI, remains a leader with models like GPT-4o and o1, offering versatile features for a wide range of tasks. Given their growing popularity, this article compares these two AI titans, focusing on performance, ease of use, stability, cost, privacy, and user feedback, ensuring all information is up-to-date and relevant to the current month.

Performance Comparison

Recent benchmarks from June 2025, as detailed in Exploding Topics: DeepSeek vs. ChatGPT, show DeepSeek R1 and OpenAI o1-1217 performing closely across several metrics. I chose them because they were released around the same time:

• DeepSeek R1 slightly outperforms in SWE-bench Verified and MATH-500, indicating strengths in coding and mathematical tasks.
• OpenAI o1 does better in MMLU, suggesting an edge in general knowledge and reasoning.

What do actual users say?

This is what one user on X said about DeepSeek and ChatGPT:

“Deepseek consistently challenges my POV whereas ChatGPT just reinforces it.”

MENAUnleashed on X, on the other hand, criticized DeepSeek, calling it “sh*t” and preferring ChatGPT for reliability.

Another X post noted that DeepSeek fabricated information, while ChatGPT was more honest about its limitations.

Ease of Use and Stability

Ease of use is crucial, and PCMag: ChatGPT vs. DeepSeek notes ChatGPT’s user-friendly interface, with additional features like custom GPTs and better web search sourcing (icons, in-text highlights, pictures).

What do actual users say about that?

DeepSeek, while powerful, can be slower and less reliable, with users on X reporting issues like “server busy” messages. For example, arthistorynews on X noted both models can make up facts for historical research, but DeepSeek was less reliable. ChatGPT also offers more platforms, including desktop apps for macOS and Windows, plus an official Chrome extension, while DeepSeek is mainly web-based with unofficial extensions available through apps like Perplexity or Hugging Face.

Cost and Accessibility

Cost is a significant factor, and DeepSeek stands out. Exploding Topics: DeepSeek vs. ChatGPT reports DeepSeek is completely free to use, with an API costing $0.07 per million input tokens and $1.10 per million output tokens, compared to ChatGPT’s GPT-4.1 at $2 per million input tokens and $8 per million output tokens. This makes DeepSeek appealing for budget-conscious users. However, PCMag: ChatGPT vs. DeepSeek notes ChatGPT’s free tier includes more features, like deep research (generating reports with 50+ sources) and image generation, with paid plans starting at $20/month for Plus and up to $200/month for Pro.

Privacy: A Growing Concern

Privacy is a hot topic, especially with DeepSeek’s recent scrutiny.

Both models collect user data by default (with opt-out options), but ChatGPT has stronger privacy policies, backed by OpenAI’s compliance with Western regulations.

DeepSeek, however, has raised alarms due to its Chinese origins and data storage practices. Chinese law allows government access to data, and DeepSeek’s privacy policy confirms user data, like AI requests and uploaded files, is stored in China.

On June 27, 2025, Germany’s data protection commissioner, Meike Kamp, urged Apple and Google to remove DeepSeek from their app stores, citing “unlawful” data transfers to China without EU-level safeguards. This follows a May 2025 warning DeepSeek ignored, prompting the ban request under GDPR. Italy and the Netherlands have also restricted DeepSeek for similar reasons, and X posts highlight growing concerns about Chinese authorities accessing user data.

ChatGPT, while not perfect, faces less regulatory heat, making it a safer bet for privacy-conscious users.

Want to test both side-by-side?

For users wanting both models, aiToggler integrates over 300 AI models, including DeepSeek and ChatGPT. It allows users to switch between models with one click and features a split chat function to compare responses side by side.

This is particularly useful for tasks where different models might excel, enabling users to choose the most appropriate answer.

Conclusion

As of June 2025, ChatGPT remains the better choice for most users due to its reliability, additional features, and better privacy practices. DeepSeek, while impressive for technical tasks and free to use, has reliability issues and privacy concerns. Given the mixed user opinions, especially around stability and ease of use, I recommend trying both to see which fits your specific needs, especially considering aiToggler for seamless integration.