Benchmark menu

Model detail

GPT-5.4 Nano

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

GPT-5.4 Nano

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

Rank
19
Rank movement
Baseline
Score
34.0
Last measured
2026-07-05
Coverage
Complete coverage

Reasoning scores

XHigh
45.6
Medium
33.6
None
23.0

Generation health

XHigh59%29%12%n 17Medium41%37%22%n 32None31%47%22%n 32 Complete coverage

Cost summary

Mean standard list-price equivalent
$0.11
Median unique-cell cost
$0.03
Unique-cell range
$0.0058–$0.39
Coverage
24 unique cells / 81 observed runs
Observed-run total
$6.14
Evidence basis
Standard list-price estimate
Mean output
58k tok
Output range
2.7k–316k tok
Token samples
81

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