Skip to content
Bencher

TestsBenchmarks · Coding

FrontierCode (Diamond)

Hardest FrontierCode subset (early version reported at Claude Fable 5 launch).

% solvedHigher is betterCounts toward:Weights: Coding ×0.5The test's websiteOfficial page ↗

The best 2 · 2 models tested

Top 2 · best setting per model · 2 models, 2 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported
FrontierCode (Diamond) leaderboard
Thinking levelSetting
1Claude Fable 5Anthropic29.3%Extra highadaptive thinking, effort=xhighMaker's own figureVendor-reportedAnthropic ↗Claude Code9 Jun 2026
2Claude Opus 4.8Anthropic13.4%Extra highadaptive thinking, effort=xhighMaker's own figureVendor-reportedAnthropic ↗Claude Code9 Jun 2026