Benchmark menu

Model detail

GPT-5.4 Mini

A static, shareable model page that remains complete without JavaScript.

GPT-5.4 Mini

Provider: OpenAI · Public slug: gpt-5-4-mini

Rank
17
Rank movement
Baseline
Score
35.3
Last measured
2026-07-05
Coverage
Complete coverage

Reasoning scores

XHigh
47.6
Medium
32.7
None
25.6

Generation health

XHigh83%17%0%n 12Medium92%4%4%n 24None79%21%0%n 24 Complete coverage

Cost summary

Mean standard list-price equivalent
$0.40
Median unique-cell cost
$0.23
Unique-cell range
$0.01–$1.52
Coverage
23 unique cells / 60 observed runs
Observed-run total
$18.43
Evidence basis
Standard list-price estimate
Mean output
66k tok
Output range
1.5k–441k tok
Token samples
60

Cost means equal-weight unique model/game/reasoning cells. Provider-account billed spend is reported separately on Charts.

Run setup

  • Overall Models: combines public reasoning variants
  • Standard GameBench code-generation prompt