Skip to content
Bencher

TestsBenchmarks · Math

ArXivMath (no tools)

MathArena final-answer problems extracted monthly from recent arXiv papers (release month noted per row).

% solvedHigher is betterThe test's websiteOfficial page ↗

The best 4 · 4 models tested

Top 4 · best setting per model · 4 models, 4 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported
ArXivMath (no tools) leaderboard
Thinking levelSetting
1Claude Opus 5.5Anthropic91.2%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—22 Sep 2026
2Claude Opus 5Anthropic90.8%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—24 Jul 2026
3Claude Sonnet 5.5Anthropic86.8%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—28 Sep 2026
4Claude Fable 5.1Anthropic82.9%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—22 Sep 2026