> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Experimental AI models from OpenAI and Google flood lmarena
- URL: https://www.testingcatalog.com/experimental-ai-models-from-openai-and-google-flood-lmarena/
- Published: 2024-11-18T08:26:09.000Z
- Updated: 2024-11-18T08:26:09.000Z
- Author: Alexey Shabanov
- Tags: Latest AI News

Recently, a bunch of new experimental models appeared on lmarena. First is the **Anonymous Chatbot**, a name previously used by OpenAI for their experimental models. This could potentially be the 4o update from November 11, as some users have reported noticing differences in model responses on ChatGPT. However, it’s also possible that this model represents something more advanced.

> anonymous-chatbot is back in the arena  
>  
> this is usually reserved for GPT-4o model updates inside ChatGPT 👀 [pic.twitter.com/OGSRJ7a6Ty](https://t.co/OGSRJ7a6Ty?ref=testingcatalog.com)
> 
> — ʟᴇɢɪᴛ (@legit\_rumors) [November 16, 2024](https://twitter.com/legit%5Frumors/status/1857827554884755553?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

In addition, Google has introduced two new models. One of them, called **Secret Chatbot**, is identified as Gemini and reportedly performs very well. This could potentially be a larger version of the [experimental model released last week](https://www.testingcatalog.com/new-ai-model-gemini-experimental-1114-debuts-on-google-ai-studio/), which Google has teased will officially arrive next week.

> Gemini-exp-1114 is now available via the Gemini API, happy building / testing! Will follow up Monday with more 🚢[https://t.co/V4PD3Dndqz](https://t.co/V4PD3Dndqz?ref=testingcatalog.com)
> 
> — Logan Kilpatrick (@OfficialLoganK) [November 15, 2024](https://twitter.com/OfficialLoganK/status/1857535825895993366?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

The second model from Google is called **Mystery Gemini 3**. Unlike **Secret Chatbot**, it does not seem to perform as impressively. Overall, Google is running a wide range of experimental models on lmarena, making it difficult to determine whether any of them stand out or represent significant advancements.

If you haven't heard of [lmarena](https://lmarena.ai/?ref=testingcatalog.com) yet, it is an LLM battleground where experimental models are being added anonymously and may only randomly appear in the Battle mode.