AI Model Reviewer

Claude Code

5 models ran in it: Claude Fable 5.1, Claude Haiku 5.5, Claude Opus 5.5, Claude Sonnet 5.5, GLM-5.3.

260 runs on 125 prompts · 80% finished

Results

The best run of each of its 125 tests first, then every other run. Filter them in the explorer

208 finished runs

Numbers: finish rate, cost, wall time, draft scores
Runs
260
on 125 prompts
Finished
80%
208 of 260
Median cost, finished run
$4.57
n=161
Median wall time, finished run
26m 43s
n=208
Recorded spend
$1,570.71
$208.71 on unfinished runs; 57 without a recorded cost

From the numbers

Set against the other harnesses with at least 3 runs (top or bottom third), and its draft-score ranks within single prompts. Nothing here is averaged across prompts.

Strengths

  • Draft score 9.5/10 on Pelican riding a bicycle (control) (Claude Opus 5.5): ranked 1st of 13
  • Draft score 9.5/10 on MacBook Pro 16" realistic 3D render (Playcode benchmark) (Claude Opus 5.5): ranked 1st of 12
  • Draft score 10/10 on Animated pelican riding a bicycle (SMIL/CSS) (Claude Opus 5.5): ranked 1st of 12
  • Draft score 9/10 on MacBook Pro 16" realistic 3D render (Playcode benchmark) (Claude Fable 5.1): ranked 2nd of 12
  • Draft score 8.5/10 on Lighthouse on a headland at night (Claude Fable 5.1): ranked 2nd of 12
  • Draft score 9.5/10 on Isometric SVG town with sprite sheet (Claude Fable 5.1): ranked 2nd of 11

Weaknesses

  • $4.57 (n=161): 9th of 12 harnesses with at least 3 (median cost of a finished run)
  • 3 runs ended as the model's own failure (DNF, class MODEL)
  • Other unfinished runs, not counted against the model: 14 HARNESS, 6 INFRA, 21 TIMEOUT, 1 STOPPED, 7 ERROR

Best and hardest run

Best run

Picked as its best draft-score rank on a prompt.

Hardest run

Picked as its costliest run that did not finish.

Draft scores by prompt

Isometric SVG town with sprite sheetClaude Fable 5.19.5/10 draft2nd of 11
Lighthouse on a headland at nightClaude Fable 5.18.5/10 draft2nd of 12
MacBook Pro 16" realistic 3D render (Playcode benchmark)Claude Opus 5.59.5/10 draft1st of 12
MacBook Pro 16" realistic 3D render (Playcode benchmark)Claude Fable 5.19/10 draft2nd of 12
Pelican riding a bicycle (control)Claude Opus 5.59.5/10 draft1st of 13
Pelican riding a bicycle (control)Claude Fable 5.18.5/10 draft9th of 13
Animated pelican riding a bicycle (SMIL/CSS)Claude Opus 5.510/10 draft1st of 12
Animated pelican riding a bicycle (SMIL/CSS)Claude Fable 5.19/10 draft6th of 12

By model

Claude Haiku 5.555100%not recorded: no finished run with a recorded cost8m 07s n=5
Claude Opus 5.512411190%$4.16 n=11128m 20s n=111
Claude Sonnet 5.5624877%$6.65 n=625m 20s n=48
Claude Fable 5.1403075%$10.36 n=3032m 25s n=30
GLM-5.3291448%$1.07 n=1421m 34s n=14
All 125 tests