Skip to main content

Enhanced Model Information Card

Feature Request

I appreciate the wide variety of models available in T3 Chat. However, it can be challenging to quickly determine which model is best suited for a specific task. While the information button beside each model provides some hints, I often find myself wondering about the optimal use-cases for newly added models.


I propose replacing the current plain text information tooltip with an interactive, stat card that appears on hover. This card could display key attributes of each model using visual indicators for example with sliders from 0 to 100.

Here are some example stats that could be implemented:

  • Speed: How fast the model processes input.

  • Reasoning: The model’s ability to perform logical or analytical tasks.

  • Writing Capability: Quality and coherence of generated text.

  • Creativity: The model’s capacity for novel or imaginative responses.

    …


I believe this enhancement would greatly improve the user experience when selecting a model.

Maybe it could also incorporate a user rating system, where users themselves can vote from 0 to 10 on how well the model performs for each attribute. Alternatively, data could be fetched from external sources such as the chatbot arena leaderboard or similar platforms.

Status: Completed11 comments

Log in to comment and vote

Comments11

  • Yellow Octopus

    Also, I heard that there might be a future feature allowing users to branch off chats and switch between models. This functionality would further simplify the process of selecting the most appropriate model for a given task. For instance, one could initially use a model with strong vision capabilities to extract data from an image, and then seamlessly switch to a model that is good for creative writing. :)

  • Lime Chocolate

    •

    Merged request

    "Fastness" scale

    Issue
    We can currently already filter by ‘Fast’ in the model picker… but unironically, all models are tagged as ‘fast’. Would be cool to have some sort of scale of how fast a model is.

    Why is this important?

    Often, I ask very simple questions that any LLM can answer me. I don’t need Claude 3.7. I just need it to be quicker than Google. And such a scale would help me weigh how long I’m willing to wait for an answer vs how complex the query is.

    More importantly, it would be good to see how long a model takes to spin up before generating an answer. Gemini 2.5 Pro regularly takes ~10 seconds before I see the first token, while Fireworks v3 is near-instant. This makes a huge difference in how… likely I am to use a model…


    Implementation Suggestion (might suck)
    Would be cool to have some sort of scale (maybe a small ‘5s‘, ‘3s’, … chip that denotes how long it takes to generate an answer to a reference question?) next to it that denotes how quick the model is?

  • Yellow Dust

    •

    Merged request

    Models Page: Context Window Size

    For each model, show the allowed context window size for each model. Add context window size to the model filtering. Since context window sizes change over time, you can use Small, Medium, Large and Extra Large instead of ‘128Kb’.

    • Amaranth Reef

      Show context size under the I in the model selection page

    • Pink Pulsar

      This is essential info! Would love to have it.

  • Sapphire Honey

    •

    Merged request

    Model Speed Indicator.

    Model speed icons to indicate how fast the model is. mostly to give DeepSeek R1 from groq its justice.

    • Bronze Fox

      Current token speed and latency would be really useful to know. Maybe some other stats from artificial analysis too, like intelligence index.

  • Sapphire Crab

    •

    Merged request

    Feature to filter by context limit.

    After the release of Llama 4 Scout with a 10M token context window, I had the following idea:

    I would like the ability to filter different models based on the amount of context they can handle. For example, it would be useful to have filters for models that support:

    • More than 100k tokens

    • More than 1M tokens

    • More than 10M tokens

    ...and so on.

    • Mark Florkowski

      Team

      Tokens are not a very “human friendly” concept, but we do have plans for how to indicate more important deciding factors to users.

      • Azure Mustard

        It might bit unintuitive at first, but it's clear that T3Chat is designed with casual and power users in mind.

        Personally, I often paste large chunks of text for LLMs to parse and summarize. A token filter would be incredibly useful for that; maybe it could be an optional toggle in settings, off by default.

        Since I'm constantly trying to balance speed, intelligence, and context limits, having a tool like this would really help me fine-tune that balance.

        Hope you'll consider adding it!

  • Coral Breeze

    •

    Merged request

    Document what model version is being used for each model

    For example, is o3-mini using o3-mini-2025-01-31 or a later version? There was an update recently on gpt-4o on their app, but is the API also using an updated version or still stuck with gpt-4o-2024-08-06?

    For advanced users it'd be nice to understand exactly which versions we're using!

    For most users this might not be relevant, but would still be cool to have it in the

    https://t3.chat/settings/models page, even if hidden behind the Show more modal.