MirrorCode (Epoch AI, with METR)

Reimplement a large existing program from its behavior alone. Built by Epoch AI with METR; score is the share of tests passed. Data: Epoch AI, CC BY.

9models

–Open Index weight

–last updated

epoch-taskspublisher

#	Model	Effort	Score	Cost / task	Harness
1	Opus 5.5	max	77.4%	–	Epoch AI
2	Fable 5.1	high	73.3%	–	Epoch AI
3	Fable 5	high	63.9%	–	Epoch AI
4	GPT-6 Astra	high	46.7%	–	Epoch AI
5	Opus 4.7	high	31.1%	–	Epoch AI
6	GPT-5.6 Sol	high	20.0%	–	Epoch AI
7	GPT-5.4	high	15.6%	–	Epoch AI
8	GPT-5.5	high	10.0%	–	Epoch AI
9	Gemini 3.1 Pro	high	8.9%	–	Epoch AI