A Reddit user conducted an experiment to see if everyday people could distinguish between various AI models. Participants were led to believe they were using a limited-time free version of ChatGPT, while in reality, they interacted with a range of local models including Qwen and Gemma variants. The experiment revealed that users noticed a decline in quality with smaller models, with complaints arising for those at the 9B parameter level and below. Participants generally preferred responses from Gemma models and found that larger models sometimes appeared to "think too much." AI
RANK_REASON This is a user-generated experiment on Reddit, not a formal research paper or industry announcement.
- AI models
- ChatGPT
- Gemma 12B
- Gemma 26B-A4B
- Gemma E2B
- Gemma E4B
- Qwen 3.5 9B
- Qwen 3.6 27B
- Qwen 3.6 35B-A3B
- Qwen 3.8 27B
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →