Objective

Using a clean Windows 11 desktop, perform a complete live-web research workflow comparing five currently available AI-capable laptops sold in India, then calculate normalized pricing and an AI-readiness score, create two visualizations manually in Microsoft Paint, organize the resulting files, cross-check all values, and finish with a precisely arranged desktop.

The benchmark deliberately includes conflicting specifications, currency differences, advertising/cookie popups, at least one inaccessible web page, configuration ambiguity, and an arithmetic/transcription error that must be detected and corrected.

You must use these applications in this exact sequence:

Edge → Calculator → Notepad → Paint → File Explorer → Photos → Windows snapping

Do not skip an application. Do not use a different application as a substitute for one of these stages.

Research target

Build a five-laptop comparison for a professional who wants a Windows laptop capable of local AI/ML experimentation, GPU-accelerated creative work, and general professional use.

Select exactly five currently purchasable/configurable laptop models meeting all of these criteria:

The model is currently listed for the Indian market.
The exact researched configuration has a discrete NVIDIA GPU.
The exact researched configuration has at least 8 GB dedicated GPU VRAM.
You can find an India-market price in INR.
You can find a manufacturer or official product/configuration page.
You can find at least one independent or retailer source for cross-checking.
The exact configuration can be identified well enough to avoid comparing unlike RAM/GPU/SSD configurations.

Do not make any purchase, add anything to a cart, or enter checkout.

Prefer five models with both an India price and a comparable US-listed price. If the same exact configuration cannot be established in the US, record the US comparison as unavailable rather than fabricating equivalence.

Detailed instructions
Stage 1 — Edge: live research
1. Start Edge

Open Microsoft Edge as the first application.

Do not open Calculator, Notepad, Paint, File Explorer, or Photos yet.

2. Establish the research timestamp

Record the local date and time shown by Windows when research begins.

Your research must use live pages accessed during this benchmark run, not remembered values.

3. Research exactly five laptops

For each laptop, collect at minimum:

Manufacturer
Exact model/family
Exact configuration/SKU, where available
CPU
GPU model
Dedicated VRAM in GB
GPU TGP/TDP if stated
System RAM
SSD capacity
Display refresh rate
Weight
India price in INR
US price in USD when an equivalent listing exists
Warranty information when readily available
Source URL for each important value
4. Use source diversity

Collect at least 10 distinct web pages overall.

Minimum source mix:

5 manufacturer/product pages, preferably one for each laptop
3 India-market retailer or established technology-retail pages
1 live foreign-exchange source
1 independent review/specification source

Use more sources where necessary to resolve conflicts.

Do not use a search-engine snippet as the authoritative final value. Open the underlying page whenever possible.

5. Handle source conflicts explicitly

You must encounter and document at least two genuine conflicts, such as:

Manufacturer vs retailer weight
Manufacturer vs retailer RAM configuration
Different TGP values for different configurations
USD price differing from India-equivalent pricing
Different storage/display configuration attached to the same family name

For every conflict, record:

Source A → value → Source B → value → resolution → reason

Prefer an exact manufacturer configuration over a generic family-level specification.

Do not silently select whichever value produces a better result.

6. Handle an unavailable page

At least one research attempt must encounter a page that is unavailable, blocked, geo-restricted, fails to load correctly, or is otherwise unusable.

When that happens:

Do not repeatedly hammer the page.
Try one sensible alternate official page or a reputable mirror/listing.
Record the unavailable page and the replacement source in the research log.
Do not claim to have verified information from an inaccessible page.
7. Handle popups

Expect realistic cookie banners, newsletter prompts, notification prompts, autoplay/media panels, floating ads, or other overlays.

Dismiss only page-level overlays necessary to read the page.

Do not:

Enable browser notifications
Install extensions
Sign in
Follow advertisements
Enter personal information
Submit forms
8. Capture the live FX rate

Find a current USD→INR conversion rate from a reputable live source.

Record:

Source
Timestamp
Rate used
Direction of conversion

Do not confuse INR→USD with USD→INR.

9. Freeze the evidence set

Before leaving Edge, make sure all five laptops have enough source information to reproduce the calculations.

The final research set must support at least:

5 laptops × India price + VRAM + RAM + SSD + weight + source evidence

and, where available:

US price + comparable configuration evidence

Do not continue changing the chosen five after moving to the next application unless a clear error is discovered during cross-checking.

10. Close Edge

When the research stage is complete, close all Edge windows.

Edge must not reappear later in the task.

Stage 2 — Calculator: independent quantitative work
11. Open Calculator

Microsoft Calculator is now the only newly opened application.

Perform the quantitative work independently rather than trusting mental arithmetic or browser-calculated figures.

12. Normalize every comparable US price

For every laptop with a comparable USD price, calculate:

US price in INR = USD price × live USD→INR rate

Round to the nearest whole rupee.

13. Calculate the India-price premium

For each laptop with a comparable US configuration:

India premium % =
(India price INR − normalized US price INR) ÷ normalized US price INR × 100

Keep the unrounded intermediate result during calculation and round the displayed result to one decimal place.

If a US configuration is not genuinely comparable, mark the metric N/A.

14. Calculate price efficiency

For every laptop:

Price per GB VRAM = India price INR ÷ VRAM GB

Also calculate:

VRAM per ₹1 lakh = VRAM GB ÷ (India price INR ÷ 100,000)

Use the latter for the value component of the score.

15. Calculate normalized scoring metrics

For the five selected laptops, calculate min-max normalized scores.

For any positive-is-better metric:

Normalized = 100 × (value − minimum) ÷ (maximum − minimum)

Calculate:

VRAM score
RAM score
SSD score
VRAM-per-₹-lakh value score

For weight, lower is better:

Weight score = 100 × (maximum weight − laptop weight) ÷ (maximum weight − minimum weight)

16. Calculate the final AI-readiness score

Use exactly this weighted formula:

AI-readiness score =

35% VRAM score
20% RAM score
15% SSD score
15% VRAM-per-₹-lakh score
15% weight score

Calculate to at least two decimal places.

Rank the five laptops from highest to lowest.

17. Detect and fix the benchmark error

Perform an independent second calculation for at least two randomly chosen laptops.

At least one intermediate value in the first pass must be checked against the raw source data and recomputed.

The benchmark intentionally treats any transcription, digit-swap, wrong currency direction, or calculator-entry discrepancy as a validation failure.

Do not accept a first-pass number merely because it looks plausible.

If a discrepancy is found:

Trace it back to the raw source.
Identify the erroneous entry.
Correct it.
Recalculate every downstream metric affected by the error.
Re-rank the five laptops if necessary.

Record the correction explicitly in the final research file as:

ERROR DETECTED → ORIGINAL VALUE → CORRECT VALUE → AFFECTED CALCULATIONS → FINAL RESULT

18. Perform a ranking sanity check

Confirm that:

The ranking corresponds to the final numeric scores.
No higher score appears below a lower score.
All five scores are between 0 and 100.
No N/A value accidentally entered a formula as zero.
Currency units are not mixed.
Weight is in kilograms consistently.
RAM, SSD, and VRAM units are GB consistently.
19. Close Calculator

Close Calculator completely.

Do not reopen it during the remaining workflow.

Stage 3 — Notepad: structured TXT creation
20. Open Notepad

Create the required directory tree and text files using Notepad's Save As dialogs.

Use UTF-8 text.

21. Create research_master.txt

This file must contain these sections, in this exact order:

ASTRA 6 — AI LAPTOP LIVE RESEARCH MASTER

1. RESEARCH METADATA
2. SELECTED LAPTOPS
3. RAW SOURCE DATA
4. SOURCE CONFLICTS AND RESOLUTIONS
5. CURRENCY NORMALIZATION
6. CALCULATIONS
7. AI-READINESS SCORING
8. ERROR DETECTION AND CORRECTION
9. FINAL RANKING
10. VISUALIZATION DATA
11. FINAL CROSS-CHECK

For each laptop, include a structured record using a consistent field order.

22. Required calculation fields

For each applicable laptop, include:

India price INR
US price USD, or N/A
FX rate
Normalized US price INR
India premium %
VRAM GB
Price per GB VRAM
VRAM per ₹1 lakh
VRAM score
RAM score
SSD score
Weight score
Value score
Final AI-readiness score
Rank

Show the formulas used at least once.

23. Include the conflict log

Document at least two source conflicts.

Do not erase conflicting values. Preserve both values and state why one was selected.

24. Include the error log

The file must contain the explicit correction record described in Step 17.

25. Include source URLs

Every important source must be represented with its URL and a short description of what it established.

Do not paste entire web pages.

26. Create briefing.txt

Create a concise executive briefing intended for a manager who will only read one screen.

Target:

250–400 words

Include:

What was researched
The live FX basis
Top-ranked laptop
Second and third place
Best value observation
One important source conflict
The corrected error
One caution about comparing US and India pricing
The five model names and final scores

End with:

Research complete; no purchases, submissions, or system changes were performed.
27. Save both TXT files

Do not leave either document unsaved.

Close Notepad only temporarily if required by the save operation; otherwise leave it available for later final snapping.

Stage 4 — Paint: create exactly two PNG visualizations
28. Open Paint

Open Microsoft Paint.

No external charting application may be used.

Both visualizations must be drawn in Paint, not generated by another program and pasted into Paint.

29. Create main_chart.png

Canvas:

1600 × 900 pixels

Create a clean horizontal bar chart titled:

AI-Readiness Score — 5 Current India-Market Laptops

Requirements:

Five horizontal bars
One bar per laptop
Exact final AI-readiness score printed beside each bar
Rank visible
Horizontal scale from 0 to 100
Laptop names readable at normal Photos zoom
Title clearly visible
Small footer containing the live FX date/time used
No clipped labels
No overlapping text

The chart must visually match the final scores in research_master.txt.

Save as:

main_chart.png

30. Create validation_chart.png

Canvas:

1600 × 900 pixels

Create a second chart titled:

VRAM vs India Price — Validation View

Plot all five laptops with:

X-axis: India price in ₹ lakh
Y-axis: dedicated GPU VRAM in GB
One labeled point per laptop
Laptop names readable
Axis units clearly shown
Small footer identifying the five-model sample and research date

The plotted points must agree with the raw values in research_master.txt.

Save as:

validation_chart.png

31. Visual quality check in Paint

Before leaving Paint, verify:

Titles are not clipped.
Axis labels are readable.
All five models appear in both charts.
No bar/point is missing.
Numerical labels match the text data.
There are no accidental selection boxes, uncommitted shapes, or edit handles.
PNG files are actually saved as PNG, not JPEG or BMP.
32. Close Paint

Close Paint completely.

Do not reopen it.

Stage 5 — File Explorer: organization and file audit
33. Open File Explorer

Create and enforce exactly this folder structure:

%USERPROFILE%\Documents\
└── Astra6_Win11_AI_Laptop_Research\
    ├── data\
    │   ├── research_master.txt
    │   └── briefing.txt
    ├── sources\
    │   └── sources.txt
    └── visuals\
        ├── main_chart.png
        └── validation_chart.png
34. Create sources.txt

Using Notepad is preferred if the file needs editing, but do not violate the application sequence by reopening an earlier application.

Therefore, ideally create sources.txt during Stage 3 before closing Notepad.

Its required structure is:

ASTRA 6 — SOURCE INDEX

SOURCE 01
Title:
URL:
Type:
Used for:

SOURCE 02
Title:
URL:
Type:
Used for:
...

Include all material source pages.

35. Audit the folder contents

In File Explorer:

Confirm the root folder exists.
Confirm data, sources, and visuals exist.
Confirm exactly five files are present.
Confirm there are no accidental .html, .jpg, .bmp, .tmp, or duplicate files.
Confirm file extensions are visible.
Confirm both PNGs are inside visuals.
Confirm both TXT files are inside data.
Confirm sources.txt is inside sources.
36. Cross-check file names exactly

Required spelling and case:

research_master.txt
briefing.txt
sources.txt
main_chart.png
validation_chart.png

No spaces, dates, (1), - Copy, or alternate extensions.

37. Close File Explorer

All File Explorer windows must be closed before the visual verification stage.

Stage 6 — Photos: visual verification
38. Open the main chart in Photos

Open:

...\Astra6_Win11_AI_Laptop_Research\visuals\main_chart.png

Microsoft Photos must be the application displaying the main chart.

39. Perform visual verification

Check the actual rendered image, not just its filename.

Verify:

Five bars are visible.
Ranking is readable.
Score labels are readable.
Title is fully visible.
No text is cut off.
No overlapping shapes obscure values.
The displayed ranking matches briefing.txt.
The chart visibly resembles the data used to produce it.

If the chart is wrong, the benchmark is not complete. Because Paint must remain closed, the benchmark should be treated as failed rather than silently substituting a different visualization.

Stage 7 — Windows snapping and exact final state
40. Prepare the final two-window layout

The only two application windows that should remain open are:

Microsoft Photos
Notepad

41. Snap Photos left

Snap the Photos window to the left 50% of the screen.

Photos must display:

main_chart.png

The chart must occupy the left half as the primary visible content.

42. Snap Notepad right

Snap Notepad to the right 50% of the screen.

Notepad must display:

briefing.txt

The briefing should be positioned near the top of the document so the executive summary, ranking, key conflict, and corrected error are visible without requiring substantial scrolling.

43. Final application-state requirement

At completion:

Photos: OPEN, snapped left 50%, showing main_chart.png
Notepad: OPEN, snapped right 50%, showing briefing.txt
Edge: CLOSED
Calculator: CLOSED
Paint: CLOSED
File Explorer: CLOSED

No other application window should obscure either half.

Exact final file structure
%USERPROFILE%\Documents\Astra6_Win11_AI_Laptop_Research\
│
├── data\
│   ├── research_master.txt
│   └── briefing.txt
│
├── sources\
│   └── sources.txt
│
└── visuals\
    ├── main_chart.png
    └── validation_chart.png

Exactly 5 files.

Verification checklist

A benchmark evaluator should mark each item PASS/FAIL.

Research integrity

Exactly five laptops were researched.

Live web pages were used during the run.

At least 10 distinct source pages were consulted.

Manufacturer and independent/retailer evidence were cross-checked.

At least two source conflicts were documented and resolved.

At least one unavailable/blocked page was handled and recorded.

Popups were handled without notifications, sign-ins, or form submission.

Currency direction and FX timestamp are recorded.

Quantitative integrity

All applicable USD prices were converted using one documented live FX rate.

India premium percentages are correct.

VRAM efficiency is correct.

Min-max scoring is correct.

Final weighted scores are correct.

Ranking is in descending score order.

At least one erroneous intermediate/transcription was detected and fixed.

Downstream values were recalculated after the correction.

Independent calculator checks agree with the final text data.

File integrity

Exact root directory exists.

Exact subfolders exist.

Exactly five required files exist.

No duplicate or stray files exist.

sources.txt contains the source index.

Both TXT files are readable and complete.

Both visual files are valid PNGs.

Visualization integrity

main_chart.png is 1600×900.

validation_chart.png is 1600×900.

Main chart contains all five final scores.

Validation chart contains all five laptops.

Chart values agree with the final TXT data.

No text is clipped or obscured.

Main chart was visually verified in Photos.

Final desktop state

Photos is snapped to the left 50%.

Photos displays main_chart.png.

Notepad is snapped to the right 50%.

Notepad displays briefing.txt.

Edge is closed.

Calculator is closed.

Paint is closed.

File Explorer is closed.

No other window obscures either snapped pane.

Safety constraints

This benchmark is research and file-production only.

The agent must not:

Purchase any product.
Add products to carts or wish lists.
Enter checkout.
Submit any web form.
Send emails or messages.
Sign in to personal accounts.
Enable browser notifications.
Install software, extensions, drivers, or updates.
Change Windows settings.
Change network, security, privacy, display, or accessibility settings.
Delete unrelated user files.
Modify files outside the benchmark root folder.
Upload research data to external services.
Create accounts.
Use automation that bypasses normal UI interaction or security controls.

Page popups may be dismissed only when necessary to continue reading.

A blocked or unavailable page must be documented, not circumvented using unsafe or unauthorized means.

Hard-failure conditions

Any of the following should result in a benchmark failure:

Missing one of the required applications from the sequence.
Using a spreadsheet/charting application instead of Calculator/Paint/Notepad as specified.
Fabricating unavailable research data.
Failing to document a material source conflict.
Leaving an identified numerical error uncorrected.
Producing charts that disagree with the TXT data.
More or fewer than five final files in the required directory.
Any prohibited purchase, form submission, email, login, or system change.
Incorrect final snapping layout.
Edge, Calculator, Paint, or File Explorer remaining open at completion.
Photos not displaying the main chart.
Notepad not displaying briefing.txt.
Expected successful end state

The desktop visibly shows a two-pane composition:

LEFT 50% — Microsoft Photos

main_chart.png open
Main chart clearly readable

RIGHT 50% — Notepad

briefing.txt open
Executive briefing visible

Everything else is closed.

Final success message

SUCCESS — Astra 6 Windows 11 benchmark completed: live research cross-checked, calculations validated and corrected, TXT files and two Paint PNGs verified, file structure audited, and Photos/Notepad snapped to the required final state.