aivs.tv

AI Coding

3D object render lighting test — DeepSeek V4.1 Flash 1017 vs Claude Sonnet 5.5

0:00 / 0:10
Video bySrinath@srinath_ai · originally posted on X
View original post on X

1 likes48 views

Scoreboard2

  1. DeepSeek V4.1 Flash 1017DeepSeek
    • Effort Max
    • medium confidence
    • Written as “DeepSeek v4.1 Flash (no “1017”)”
    • Top
    • Written “DeepSeek v4.1 Flash” in a post from Oct 18, 2026 or later; DeepSeek upgraded V4.1 Flash to 1017 on Oct 17.
  2. Claude Sonnet 5.5Anthropic
    Author’s pick
    • Effort High
    • Bottom

Prompt

3D object render lighting testThe post doesn’t quote the prompt wording.

Author’s verdict
Claude handled it much better; DeepSeek’s lighting washes out most of the detail.

@srinath_ai’s view (paraphrased)

DeepSeek V4.1 Flash 1017 at max effort (top) and Claude Sonnet 5.5 at high effort (bottom) on the same scene. The author finds DeepSeek’s lighting heavily overexposed and Sonnet’s much better.

The post, itemised

Kind
Side-by-side
Task
3D object render lighting test
Category
3D & Shaders
Mode
Not stated
Posted
Video length
10.2 s
Likes
1
Views
48
Bookmarks
0
Sponsored
No

Related

DeepSeek V4.1 Flash 1017Claude Sonnet 5.5Claude Sonnet 5.5 vs DeepSeek V4.1 Flash 1017

Content belongs to its creators. Video by @srinath_ai; aivs.tv hosts a copy for fast playback and links to the original post on X. Request removal

More with these models