Which LLM should you use? Token Monster automatically combines multiple models and tools for you

TokenA new AI Chatbot platform has launched its Alpha overview, aimed at modifying how users interact with large languages ​​(LLM) models.

Developed by Matt Shumer, co-founder and CEO of Othersideai and his Hit ai Writing Assistant Hyperwrite AiThe key point of sale of Token Monster is its ability to transport user prompts to the best LLM available for the task to be accomplished, offering improved outings by taking advantage of the forces of several models.

There are seven major LLM currently available via a symbolic monster. Once a user hits something in the entrance box has been invited, Token Monster uses pre-prosperies developed by iteration by Shumer himself to automatically analyze the user’s entry, decide which combination of several available models and linked tools is best suited to respond, then provides a combined response by taking advantage of the forces of these models. LLM available include:

• Anthropic Claude 3.5 Sonnet
• Anthropic Claude 3.5 opus
• OPENAI GPT-4.1
• OPENAI GPT-4O
• Perplexity ai PPLX (for research)
• OPENAI O3 (for reasoning)
• Google Gemini 2.5 Pro

Unlike other chatbot platforms, Token Monster automatically identifies which LLM is the best for specific tasks – as well as tools connected to LLM, such as web search or coding environments – and orchestra a multi -model workflow.

“We just build the connectors to everything, then a system that decides what to use when,” said Shumer.

For example, he could use Claude for creativity, O3 for reasoning and PPLX for research, among others. This approach eliminates the need for users to manually choose the right model for each prompt, simplifying the process for anyone who wants high quality custom results.

Operation of the elements

The Alpha preview, which is currently free to register for Tokenmonster.ai, allows users to download a range of file types, including Excel, PowerPoint and Docs.

It also includes features such as extraction of the web page, persistent conversation sessions and a “fast mode” which automatically rolls towards the best model without user input.

At the heart of Token Monster is open, a third -party service that acts as a bridge towards several LLM, and in which Shumer has invested a small sum, by admission.

This architecture allows Token Monster to draw from a range of models of different suppliers without having to build separate integrations for each.

Price and availability

Currently, Token Monster does not charge stable monthly costs.

Instead, users only pay for the tokens they consume via OpenRouter, which makes it flexible for different levels of use.

According to Shumer, this model was inspired by Cline, a tool that allows high expenditure users to access an unlimited AI power, allowing them to obtain better outings by simply using more calculation resources.

Workflows in several stages produce rhic LLM responses

The AI ​​workflows of Token Monster extend beyond the simple quick routing.

In an example, the chatbot can start with a research phase using web research APIs, transmit this data to O3 to identify information gaps, then create a outline with Gemini 2.5 Pro, a text project with Claude Opus and refine it with Claude 3.5 Sonnet.

This orchestration in several steps is designed to provide richer and more complete responses than a single LLM could generate alone.

The platform also includes the possibility of saving sessions, with data stored safely using the open source database service Supabase. This ensures that users can return to current projects without losing their work, while giving them control of the recorded data and what is ephemeral.

A non -traditional CEO

In a notable experience, Token Monster’s leadership was given to the Claude d’Anthropic model.

Shumer announced that he was determined to follow each decision taken by “CEO Claude”, calling him a test to see if an AI can effectively manage a business.

“Either we have revolutionized management forever or made a huge mistake,” he wrote on X.

Emerging from the controversy of the 70-B reflection

The launch of Token Monster comes less than a year after Shumer faced the controversy over its launch and its ultimate retraction of the 70B reflection, A refined version of Meta’s Llama 3.1 which was Initially presented as the most efficient open source model in the worldbut which quickly became subject to criticism and Fraud accusations After third -party researchers could not reproduce its declared performance on third -party reference tests.

Shumer apologized And said the problems were born from errors made due to speed. The episode highlighted the challenges and risks of the rapid development of AI and the importance of transparency in the versions of the model.

Upcoming MCP integrations

Shumer said that his team on the token monster also explores new capacities, such as integration with model context protocol servers (MCP) which allow websites and companies to ensure that LLM use their knowledge, tools and products to obtain superior tasks than the generation of text or image.

This would allow Token Monster to connect with the internal data and services of a user, opening the possibilities so that it manages tasks such as the management of customer support tickets or interfacing with other commercial systems.

Shumer stressed that the token monster is still very good in its early stages. Although it already supports a series of powerful features, the platform remains an alpha product and should see quick iterations and updates because more and more users provide comments. “We will continue to iteration and add things,” he said.

A promising experience

For users who wish to take advantage of the combined power of several LLM without the harassment of the model switching, the token monster could be an attractive choice. It is designed to operate for people who do not want to spend hours adjusting the prompts or testing different models themselves, allowing the automated routing of the system and the workflows in several steps to manage complexity.

As the abilities of Token Monster increase, it will be interesting to see how users and companies adopt it – and how its experience with management led by AI. For the moment, it is a promising addition to the rapid expansion landscape of AI chatbots and digital assistants.