Harness comparisonAI Coding

Horse show-jumping, Muse Spark 1.3 in 3 harnesses — OpenAI Codex vs Claude Code vs DeepSeek Harness

0:00 / 0:16
DeepSeek Harness
Posted byLatent Operator@LatentOperatorX · on X
View post on X

14 likes2,253 views

Scoreboard3

  1. Time
    15.3 min
    Cost
    $1.05
    Tokens
    73.1K
    • Effort high
  2. Time
    21.6 min
    Cost
    $1.18
    Tokens
    163.5K
    • Effort high
  3. Time
    10.1 min
    Cost
    $1.21
    Tokens
    187K
    • Effort high

Prompt

Horse show-jumping, Muse Spark 1.3 in 3 harnessesThe post doesn’t quote the prompt wording.

Muse Spark 1.3 at high reasoning gets the same prompt for a 3D horse show-jumping animation in Three.js, run in Codex, Claude Code and DeepSeek Harness. Round 3 of the poster’s Harness Cup.

Same model in every harness: Muse Spark 1.3. More Claude Code vs OpenAI Codex. More OpenAI Codex vs DeepSeek Harness. More Claude Code vs DeepSeek Harness. All harness comparisons.

The post, itemised

Task
Horse show-jumping, Muse Spark 1.3 in 3 harnesses
Mode
Not stated
Posted
Video length
16 s
Likes
14
Views
2,253
Bookmarks
3
Sponsored
No

NotesOnly the harness changes; the post’s scores beyond Codex’s 6.3/10 are cut off in the text.

Cost & time, as stated in the postAs shown on screen: Codex ~$1.05, 15 min 16 s, 73.1k tokens (6.3/10 per the poster’s AI judge); Claude Code ~$1.18, 21 min 37 s, 163.5k tokens; DeepSeek Harness ~$1.21, 10 min 8 s, 187k tokens.

Related

Posted by @LatentOperatorX on X. aivs.tv claims no ownership: it hosts a copy for fast playback and links to the post on X. Request removal

More with these models