• MODEL LIST
  • BENCHMARKS
  • SELECTION
  • HUMAN

How to compare LLM performance with your AI

Connect Baidu Qianfan to your AI client to run side-by-side tests across different model versions.

Ask AI about this page

Short answer

How can I compare different Ernie models?

You can use your AI to pull a list of available Ernie models and then send specific prompts to each one to compare their outputs. This lets you see how different versions handle your unique logic or creative requirements.

Testing outcomes

What you'll discover

The results of your model evaluation.

MODEL LIST

Available model inventory

Your AI identifies every Ernie version currently accessible through the connection. You'll know exactly which specific versions you can target for testing.

BENCHMARKS

Performance comparison

By running the same prompt through multiple models, you get a direct comparison of how each version handles your instructions. This highlights differences in reasoning or style.

SELECTION

The best model fit

The AI helps you narrow down which specific model version provides the most accurate or useful responses for your use case. You'll stop guessing which version to use.

HUMAN

Final decision making

The AI provides the raw outputs and model lists, but you decide which model's behavior meets your quality standards. You make the final call on which version to deploy.

The workflow

What your AI does when the task arrives.

The AI handles the technical calls to the Baidu Qianfan API so you can focus on the results.

  1. Inventory check

    Your AI checks the Baidu Qianfan platform to see which Ernie models are ready for use. This ensures you aren't trying to test a version that isn't available.

    list_models

  2. Model selection

    You tell your AI which specific models from the list you want to compare. The AI prepares to send instructions to each one.

    list_models

  3. Execution

    The AI sends your test prompts to the selected Ernie models. It manages the communication for each individual request.

    chat_completions

  4. Output analysis

    The AI gathers the responses from every model version you tested. It presents them clearly so you can compare the text side-by-side.

    chat_completions

Try it

Copy these to start.

Use these prompts to begin your evaluation.

Starting points

These are starting points. Swap out the specific task or model names to match what you actually need to test.

Baidu Qianfan Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_nSPMSdk1YEMCwy7aSAehImUREP7520Z8lSHvuSUL/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Baidu Qianfan capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "baidu-qianfan-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_nSPMSdk1YEMCwy7aSAehImUREP7520Z8lSHvuSUL/mcp"
    }
  }
}
Copy into chat04

  • List all available Ernie models so I can see what's available for testing.

  • Run this prompt through the Ernie-4.0 model and the Ernie-3.5 model to compare the results: [Insert your prompt here]

  • Compare how different Ernie versions handle this specific JSON formatting task: [Insert your task here]

  • Which Ernie models are currently available in my Baidu Qianfan account?

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Baidu Qianfan once, then ask.

Just link your account once. Your credentials stay encrypted and secure, and you can immediately start testing models through your preferred AI client.

Connect Baidu Qianfan to your AI

FAQ

How this task behaves

  • 01

    Can the AI automatically pick the best model for me?

    No. The AI provides the outputs and the model list, but you must evaluate the results and decide which model performs best for your needs.

  • 02

    Can I use this to change my model settings?

    No. The AI can only list models and send chat completions. It cannot modify your account settings or model configurations.

  • 03

    Can I test models that aren't in the list?

    No. The AI can only interact with models that are returned by the list_models tool.

  • 04

    Will the AI save these test results for me?

    No. The AI performs the comparison in your current chat session. You are responsible for saving any outputs you want to keep.

  • 05

    Does the AI handle the API keys?

    Yes. Once you connect, the AI uses the connection managed by Vinkius to make the calls.

  • More questions about Baidu Qianfan? The Connector page answers them. See everything the Baidu Qianfan Connector can do

Connect Baidu Qianfan to Claude, Cursor, ChatGPT & more