MODELOBSERVATORY

Comparative study / C-009

Project 09

One numbered brief. Independent constructions and project names. Inspect the outputs and process records without collapsing their differences.

Select outputs
A
SolKestrel 7
gpt-5.6-sol
B
TerraKestrel 7
gpt-5.6-terra

Numbered brief context

Create a new folder for this project 9. Create a high-fidelity, first-person 3D spaceship walkthrough game named "Kestrel-7" inside a single index.html file using HTML, CSS, Three.js, and PointerLockControls from a CDN. Target User: Fans of atmospheric, narrative-driven retro-…

This group aligns by numbered project brief. Archived excerpts and chosen project titles may differ between model sessions.

Efficiency ledger

Session metrics at a glance

Provider-reported totals shown on one scale. Reasoning is listed separately and is not added to total tokens.

SolKestrel 7
gpt-5.6-sol
Total tokens
633.3K
Total cost
$1.0922
Elapsed
9m 50.505s
Active estimate
≈ 9m 51s
API calls
26
Input / context
87.3K
Output
13K
Cache R/W
533K / 0
Reasoning reported
1,668

Cost · input $0.4367 / output $0.3891 / cache $0.2665

Skills loaded
  • ui-ux-pro-max
TerraKestrel 7
gpt-5.6-terra
Total tokens
476.8K
Total cost
$0.4414
Elapsed
7m 39.378s
Active estimate
≈ 7m 39s
API calls
22
Input / context
74.7K
Output
10.5K
Cache R/W
391.7K / 0
Reasoning reported
2,025

Cost · input $0.1866 / output $0.1569 / cache $0.0979

Skills loaded
  • ui-ux-pro-max

Stage controls

Select outputs to inspect

On narrow screens, select 1 model at a time to keep each workspace focused.

Comparative viewing

Outputs remain dormant until selected.

Open 1 model for focused inspection, or use a wide screen to stage both workspaces side by side. Live previews still initialize only on request.