Model detail
GPT-5.4
Provider: OpenAI · Public slug: gpt-5-4
- Rank
- 2
- Rank movement
- Baseline
- Score
- 55.0
- Last measured
- 2026-07-05
- Coverage
- Complete coverage
Reasoning scores
- XHigh
- 66.7
- Medium
- 49.1
- None
- 49.3
Generation health
XHigh88%12%0%n 8Medium88%12%0%n 8None75%25%0%n 8 Complete coverageCost summary
- Mean standard list-price equivalent
- $1.08
- Median unique-cell cost
- $0.51
- Unique-cell range
- $0.07–$3.29
- Coverage
- 24 unique cells / 24 observed runs
- Observed-run total
- $25.88
- Evidence basis
- Standard list-price estimate
- Mean output
- 69k tok
- Output range
- 3.3k–218k tok
- Token samples
- 24
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