Claude Fable 5 (high)

anthropic/claude-fable-5

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	65	2041.3	$2.10882	$0.001033	rating point	2026-06-09
Reason	ProofBench	10	0.77	$1.63546	$2.124	correct answer	2026-06-09
Agent	DeepSWE	19	0.596	$3.75787	$6.307	solved task	2026-06-09
Agent	DeepSWE	28	0.654	$6.08819	$9.314	solved task	2026-06-09
Agent	DeepSWE	34	0.686	$9.17764	$13.38	solved task	2026-06-09
Agent	DeepSWE	39	0.699	$13.4145	$19.19	solved task	2026-06-09
Agent	DeepSWE	43	0.697	$21.6347	$31.03	solved task	2026-06-09

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