Inference
Epoch AI

MirrorCode

Evaluated by Epoch AI on their own harness.

ModelsModels with a result.8
Top resultBest result on this test.N/A
Top-3 spreadPoints from first to third.N/A
YearYear of release.N/A

Result and price

020406080100$3$10$30$100RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

8 results
#ModelResultThe score from the source.
—Claude Fable 5.1high effort73.3%
—Claude Fable 5high effort63.9% ±10.4
—GPT-6 Astrahigh effort46.7%
—Claude Opus 4.7high effort31.1% ±9.0
—GPT-5.6 Solhigh effort20.0% ±9.2
—GPT-5.4high effort15.6% ±7.7
—GPT-5.5high effort10.0% ±6.0
—Gemini 3.1 Prohigh effort8.9% ±4.6