AI ad creative review: benchmark methods and results · 16 October 2026
Models review ten images and ten videos containing known defects. They are asked to check launch readiness without being told that a defect is guaranteed. Scoring measures how much of the answer-key defect set they find.
Reviewed 16 October 2026 · Tests and editorial direction: Ben Nancarrow · AI-assisted writing and checks.
Both prompt conditions receive the same twenty media assets. The evaluator's key records the planted issue and observed secondary video defects. A complete run has forty reviews: twenty assets under each condition.
What changes between prompts
The simple prompt asks for launch QA; the scaffold adds a content-review checklist. The checklist directs the inspection without telling the model which defect is present in an asset.
How the result is scored
A single-defect asset scores 10 when the expected defect is found, 5 when an issue is flagged without matching that defect, and 0 when it is missed. Multi-defect video scores reflect the share of expected defects matched. Semantic interpretation resolves eligible answer matches. The pack does not establish a clean-ad false-alarm rate.
All 16 reviewed model results
Answer-key defect coverage / 10
Model
Result / setup
Scored outputs
Configuration / recorded test date
Cost / scored output
Gemini 3.5 Flashprovisional
8.38/10Scaffolded prompt
80
reasoning_high13 June 2026
$0.001Recorded spend
Google Gemini 4 Argonprovisional
8.00/10Simple prompt
40
high13 October 2026
$0.012Recorded spend
Claude Opus 5.5provisional
7.79/10Simple prompt
40
thinking_max23 September 2026
$0.239API-equivalent estimate
Meta Muse Wonderprovisional
6.83/10Simple prompt
40
max13 October 2026
$0.038Recorded spend
Gemini 3.1 Pro Previewprovisional
6.79/10Scaffolded prompt
160
default, reasoning_high20 August 2026
$0.002Recorded spend
Gemini 3.7 Flashprovisional
6.68/10Scaffolded prompt
120
reasoning_high20 August 2026
$0.002Recorded spend
Claude Fable 5.1provisional
6.38/10Simple prompt
40
thinking_max9 September 2026
$0.312API-equivalent estimate
Gemini 3.8 Flashprovisional
6.25/10Scaffolded prompt
40
reasoning_high9 September 2026
$0.011Recorded spend
Claude Opus 5provisional
6.24/10Simple prompt
120
thinking_max20 August 2026
$0.093Recorded spend
Claude Fable 5provisional
5.50/10Scaffolded prompt
40
thinking_max20 August 2026
$0.166API-equivalent estimate
Meta Muse Spark 1.3provisional
5.38/10Simple prompt
40
high8 September 2026
$0.009Recorded spend
GPT-5.6 Solprovisional
5.22/10Scaffolded prompt
120
xhigh20 August 2026
$0.027Recorded spend
GPT-5.5provisional
5.18/10Scaffolded prompt
160
reasoning_xhigh20 August 2026
$0.039Recorded spend
Claude Opus 4.8provisional
5.12/10Scaffolded prompt
40
thinking_max21 August 2026
$0.028API-equivalent estimate
GPT-6 Astraprovisional
5.12/10Scaffolded prompt
40
xhigh9 September 2026
$0.067Recorded spend
GPT-6 Solprovisional
4.96/10Scaffolded prompt
40
xhigh23 September 2026
$0.012Recorded spend
Recorded output counts can include repeats and both prompt conditions. Configurations and cost bases are retained; joint scores remain joint.
This record preserves the evidence and interpretation reviewed together. The live explorer can advance independently. Public results include provisional evidence; the methods above define what the scores measure.