Using the current AI coding assistant, the responses consistently hallucinate methods.
I found this to be a problem when using regular ChatGPT v4, but this went away when i moved to v5.
If you are not using it already, I suggest moving your AI model to V5. I realize it is more expensive, but it at least useful!
1 Answer, 1 is accepted
Hello, Ian,
Thank you for sharing your experience with the AI Coding Assistant and your observations regarding model accuracy.
The AI Coding Assistant for Telerik UI for WinForms is designed to provide code suggestions using specialized context relevant to Telerik components.
At present, the AI assistant may use models such as ChatGPT v4, Claude Sonnet 4, or GPT-5, depending on your own preferences. Claude Sonnet 4 and GPT-5 are the most advanced LLM models at the moment, hence they are optimal for code generation and reduced hallucinations. Upgrading to GPT-5 or Claude Sonnet 4 has shown significant improvements in response accuracy during testing.
Your suggestion to move to GPT-5 is highly appreciated. Thank you for sharing it with the community. We are continually working to enhance the AI Coding Assistant’s reliability and effectiveness, trying to provide you with ready-to-use solutions faster than before.
Regards,
Nadya | Tech Support Engineer
Progress Telerik
Love the Telerik and Kendo UI products and believe more people should try them? Invite a fellow developer to become a Progress customer and each of you can get a $50 Amazon gift voucher.
Rank 2
Bronze
Bronze
Iron
Yes, you can use GPT-5 with the AI Coding Assistant.
Regards,
Nadya