---
title: "Manus tested on research tasks"
canonical: "https://agenticfeats.com/c/manus-tested-on-research-tasks/"
agents: ["Manus"]
evidence_grade: "B"
outcome: "partial"
job_type: "Research"
roles: ["Marketing"]
source_url: "https://technologyreview.com/2025/03/11/1113133/manus-ai-review"
source_date: "2025-03-11"
published: "2026-10-13"
updated: "2026-10-13"
---

> From Agentic Feats: sourced examples of AI agents at work. Map of the site for agents: https://agenticfeats.com/llms.txt

# Manus tested on research tasks

Manus researched China tech reporters, NYC two-bedroom listings, and Innovators Under 35 candidates for an MIT Technology Review reporter, who steered it with feedback.

Result: It produced a 30-journalist list, a tiered NYC apartment list in under half an hour, and only five solid nominees out of fifty requested.

## How it was set up

Manus general agent with invite access. Uses Claude 3.5 Sonnet and fine-tuned Qwen models. The author watched the Manus's Computer window and took over for captchas.

## Worth knowing

Frequent crashes and server overload, paywall and captcha blocks, and a three-hour nomination task that stalled. First answers were lazy until the author pushed back.

## Prompt

```text
Compile a list of [number] notable [type of person] covering [topic]. For each, give their current outlet and notable work. Be consistent and thorough, and check for recent employer changes.
```

## Evidence note

Published hands-on review with task-by-task outcomes and timings; no output artifacts linked

## Sources

- Original: https://technologyreview.com/2025/03/11/1113133/manus-ai-review

## The receipt

- Agent: Manus
- Evidence: Numbers (grade B, a specific number stated)
- Outcome: Partial
- Job type: Research
- Roles: Marketing
- Source: Blog
- Published: 2025-03-11
- Added: 2026-10-13