Home > Forum > General > AI agent - current version

AI agent - current version
0

12.03.2026
11:18
Michal Leszczynski (Raben)

Hello.

Do you know which models are used by Webcon when AI prompt is used?
- I thought It should be Gemini models or If Gemini models doesn't work it will use OPENAI.

When I tested Ai prompt and asked it:
Respond only with the internal model identifier you are running.
Do not explain anything.
Response:
gpt-3.5-turbo

I added to prompt. Use the highest model.
Then response was:
gpt-4-turbo-2024-05-13

I checked it version at 2025.2.1.259 and 2026.1.3.109. The same answer.
This version are pretty old and slow.
I read that you can change model in new version 2026, but I can find this option.

Reply  
12.03.2026
16:21
WEBCON
Adrian Baszak (WEBCON)

Hi Michał,

AI PROMPT currently uses the Gemini 2.0 flash-lite model.

You saw a different model mentioned in the response because chat models cannot check their own version. If they mention one, they are usually repeating something from the system prompt or generating a plausible guess based on training patterns.
In AI PROMPT no system prompt is added automatically, so in this case it is the latter. The model only receives what is explicitly provided in the arguments.

AI Agents use the Gemini 2.5 flash-lite model, but this information is not included in their system instruction.

Changing the model version for each AI functionality is possible by configuring the self-hosted AI PROXY.

See:
https://docs.webcon.com/docs/2026R1/Studio/SystemSettings/GlobalParams/#4-ai-settings
https://docs.webcon.com/docs/2026R1/Guides/AIProxy
https://docs.webcon.com/docs/2026R1/Guides/AIProxy/AIConfiguration#methodtypesconfiguration

Reply  
Solution (1) Helpful (0) Wrong answer (0)
Menu
Czerwone Maki 87
30-392 Krakow, Poland
office@webcon.com
+48 12 443 13 90