GPT-5.4 (none)

openai/gpt-5.4-20260305

Every ranked configuration of this model. Rank is dollars per unit of score on that job, and the unit is the job's own — the `per` column names it, because a rating point and a solved task are not the same thing to divide by. Displayed values are rounded; exact values in site.json and the dumps.
job	benchmark	#	score	task $	ratio $	per	released
Code	ALE-Bench	18	1086.03	$0.052314	$0.00004817	rating point	2026-03-05
Code	ALE-Bench	36	1520.72	$0.26258	$0.0001727	rating point	2026-03-05
Code	ALE-Bench	54	1607	$0.612174	$0.0003809	rating point	2026-03-05
Reason	ProofBench	29	0.56	$3.20201	$5.718	correct answer	2026-03-05
Agent	DeepSWE	31	0.518	$5.65246	$10.92	solved task	2026-03-05

These are the numbers on #frontier. How a fixed capability's cost falls over time is on /methodology.