DEDUPED Hot AI repos, turned into use cases Get the Monday 5

How to compare planner systems on MobilePA-Bench by category with Qwen Planner Agent

Read the Overall and category scores to shortlist planner systems for a phone automation project.

Short answer: Open the Performance section of the README in the Tongyi-MAI/Qwen-Planner-Agent repository; then Compare the Overall column with the Tool Use, Memory, Skills, and Sub-agent columns for the systems you care about; read the footnote that says cost covers output tokens only, excluding input tokens, external tool charges, device execution, and extra Harness processing. All steps.

  • Research
  • Beginner
  • 15 minutes
  • From the README
  • Any agent

Who it is forA product lead or analyst choosing which planner model to evaluate further for a mobile task automation product.

The problemA single overall score hides differences in tool use, memory, Skills, and sub-agent coordination.

The outcomeA shortlist of systems ranked by the category that matters for your use case, with the cost caveat noted.

What you need first

  • A web browser to open the project website and the arXiv page
  • A BibTeX file or reference manager if you plan to cite the paper

Steps

  1. Open the Performance section of the README in the Tongyi-MAI/Qwen-Planner-Agent repository.
  2. Compare the Overall column with the Tool Use, Memory, Skills, and Sub-agent columns for the systems you care about.
  3. Read the footnote that says cost covers output tokens only, excluding input tokens, external tool charges, device execution, and extra Harness processing.
  4. Record the systems that lead in your priority category and mark them for a separate evaluation.

Prompt for your agent

Read the Performance table in the README of the Tongyi-MAI/Qwen-Planner-Agent repository and produce a ranked list of systems for <priority category: Tool Use, Memory, Skills, or Sub-agent>. Expected output: a table with system name, the priority score, and the Overall score. Verify that the scores match the README values and that you state the cost exclusions.

Caveat. The scores are the authors' reported results on MobilePA-Bench, and the repo does not include the benchmark or code to reproduce them.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for Tongyi-MAI/Qwen-Planner-Agent and "Compare planner systems on MobilePA-Bench by category"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer