ChatGPT vs Gemini: A blog page people finish: 46 KB vs 4.4 MB
Not a landing page — a page of prose, with contents, progress and what to read next. One entry came back at 46 KB, the other at 4.4 MB. Both are here, unedited.
GPT 5.6 Sol (Extra High)vsGemini 3.6 Flash 5 (High)

Too close to call
GPT 5.6 Sol (Extra High)
- Technical score
- 100
- Web app completion time
- 7m 0s
- Requirements met
- 7/7
Gemini 3.6 Flash 5 (High)
- Technical score
- 86
- Web app completion time
- 9m 0s
- Requirements met
- 6/7
Which one did it better?
No votes yet — be first
What separated them?Websites scorecard · answer any of them
- Which page is easier to read?
- Which page feels more finished?
- Which one followed the brief more closely?
Tap your answer again to take it back. Votes are kept apart from the automated checks and never quietly averaged into them. How scoring works.
Word for word, as both models received it
I want you to create a blog post page using HTML. Our goal is to make the blog posts easy to read and satisfying for visitors. The other posts on the site and the internal links should be placed in locations that encourage users to continue browsing the site. I’ve placed a sample blog post in the folder.
Scroll inside the box for the rest of the brief.
Requirements met
| Requirement | ChatGPT | Gemini |
|---|---|---|
| No external assets, libraries or CDNs | Met | Not metRequests fonts.googleapis.com |
| Actually draws something when opened | Met | Met |
| Links on to other posts on the same site | Met | Met |
| A contents list for the article | Met | Met |
| Reading progress the reader can see | Met | Met |
| Related posts at the end | Met | Met |
| Semantic structure rather than a wall of divs | Met | Met |
Automated checks
GPT 5.6 Sol (Extra High)
GPT 5.6 Sol (Extra High) · web app
- Total size
- 44.6 KB
- Painted area
- 31.1%
- Files produced
- 5
- Outside requests
- None
- Technical score
- 100
Gemini 3.6 Flash 5 (High)
Gemini 3.6 Flash 5 (High) · web app
- Total size
- 4.2 MB
- Painted area
- 6.2%
- Files produced
- 8
- Outside requests
- fonts.googleapis.comfailed
- Technical score
- 86
Timings are measured in each provider’s consumer web app, so they include network and interface conditions — they are not model inference latency.








