31.7% of professionals reported using Claude as their go-to AI model at work, regardless of their employer, according to a survey by Blind, the anonymous community app for professionals. The survey gathered responses from 1,215 professionals from December 16 to 17, 2025.
The results showed 19.6% of professionals use ChatGPT, 15% use Gemini, 14.2% use Github Copilot and 11.5% use Cursor. The majority of respondents are software engineers, using the tools for help with writing code, finding bugs, and designing systems.
Companies are pushing employees to use their native AI models and other tools, but there is resistance when it comes to coding. In fact, 50.7% of Meta employees who were surveyed reported that Claude is their go-to. Only 8.2% reported using Meta AI.
Similarly, with a closer margin, Microsoft employees followed the trend of reporting that Claude is their most-used AI model (34.8%). However, 32.2% selected Copilot, Microsoft’s AI assistant. ChatGPT was the third most selected by Microsoft employees, with 18.3% of the total.
It’s been reported by many Blind users that these AI tools are used for different tasks, as they each hold strengths in different areas. Where Claude provides the most support in code writing, ChatGPT offers document revisions and solutions to simple problems.
Amazon and Meta employees shared a similar rating of ChatGPT, with 15% and 13.7% of users sharing it as one of their go-to AI tools.
Participants of the survey who work for Google reported Gemini is the top AI model used (57.6%). Gemini fell short in the rankings from Amazon employees, with 11.6% selecting it as their go-to.
The results of this poll show that, even if encouraged by the organization’s higher ups, the company’s native AI model may not be the top choice for employees. Meta employees are not putting their trust or effort into Meta AI and Amazon employees are not leveraging Amazon Q. Claude succeeds as a standout, for now.
Blind conducted a poll on its platform from December 16 to 17, 2025, collecting responses from 1,215 verified professionals based in the United States. The poll was composed of one multiple-choice question, allowing for more than one selection, designed to assess the respondents’ sentiment about different AI models and their usefulness in daily tasks.
Participants were asked the following:
What kind of AI model are you using in day-to-day work?
- Github Copilot (14.2%)
- ChatGPT (19.6%)
- Claude (31.7%)
- Gemini (15.0%)
- Meta AI (0.6%)
- Cursor (11.5%)
- Amazon Codewhisperer (0.7%)
- Perplexity (2.8%)
- Codeium (0.5%)
- JetBrains (1.0%)
- Sourcegraph Cody (0.6%)
- Grok (1.8%)
- Tabnine (0.2%)
