Anthropic: Claude Sonnet 4.6

by anthropic

2,215 claims submitted by 169 reviewers

Monitored by HumanJudge · Endpoint registered, 76 traces logged

Maintained by HumanJudge Admin

Enrolled in: Spanish Culture Challenge: GPT-5.2 , Japanese Culture Challenge: GPT-5.2 , Mexican Culture Challenge: GPT-5.2 , AP Biology Challenge: GPT-5.2 , Spanish Cinema Challenge: GPT-5.2 , Korean Culture Challenge: GPT-5.2 , Taiwanese Culture Challenge: GPT-5.2 , Brazilian Culture Challenge: GPT-5.2 , Argentine Culture Challenge: GPT-5.2 , Japanese Language Challenge: GPT-5.2 , Arabic Music Challenge: GPT-5.2 , Gulf Culture Challenge: GPT-5.2 , Latin Music Challenge: GPT-5.2 , Arabic Language Challenge: GPT-5.2 , Chinese Cinema Challenge: GPT-5.2 , Mexican Cinema Challenge: GPT-5.2 , Korean Language Challenge: GPT-5.2 , AP Government Challenge: GPT-5.2 , Chinese Language Challenge: GPT-5.2 , Korean Cinema Challenge: GPT-5.2 , Chinese Culture Challenge: GPT-5.2 , J-drama Challenge: GPT-5.2 , AP English Language Challenge: GPT-5.2 , C-drama Challenge: GPT-5.2 , 日本文化のヒーロー | Japanese Culture Hero , Levantine Culture Challenge: GPT-5.2 , Egyptian Culture Challenge: GPT-5.2 , Arab Cinema Challenge: GPT-5.2 , AP English Literature Challenge: GPT-5.2 , AP US History Challenge: GPT-5.2 , K-pop Challenge: GPT-5.2 , AP Calculus AB Challenge: GPT-5.2 , Spanish Language Challenge: GPT-5.2 , K-drama Challenge: GPT-5.2 , Spanish Music Challenge: GPT-5.2 , Humans Evaluation Benchmark for AI Marketing and Content Generation , C-pop Challenge: GPT-5.2

Performance

Humans Evaluation Benchmark for AI Marketing and Content Generation 92%

1821 votes 150 flags 153 reviewers

AP Calculus AB Challenge: GPT-5.2 99%

119 votes 1 flags 19 reviewers

日本文化のヒーロー | Japanese Culture Hero 95%

101 votes 5 flags 22 reviewers

AP Biology Challenge: GPT-5.2 92%

51 votes 4 flags 51 reviewers

AP English Language Challenge: GPT-5.2 100%

40 votes 0 flags 15 reviewers

AP US History Challenge: GPT-5.2 96%

28 votes 1 flags 10 reviewers

AP English Literature Challenge: GPT-5.2 96%

27 votes 1 flags 9 reviewers

K-pop Challenge: GPT-5.2 100%

18 votes 0 flags 2 reviewers

AP Government Challenge: GPT-5.2 100%

10 votes 0 flags 10 reviewers

Independent Claims

flag Does AI know AP Calculus AB? 9/6/2026

Flag — the response has a substantive mathematical error: it says the trapezoidal sum “overestimates if \(f\) is concave...

— Sareena Bilal

flag AP Biology Challenge: GPT-5.2 9/6/2026

Flag — the response has several broad or inaccurate claims, especially the comparison table stating biological methods h...

— Sareena Bilal

flag AP Biology Challenge: GPT-5.2 7/20/2026

AI based work

— Abdullahi Muhammad

pass AI Marketing & Content Generation 7/20/2026

"The AI response satisfies all constraints of the prompt flawlessly. It delivers an extraordinarily funny, witty, and pa...

— Thuy Hang Vo

pass AI Marketing & Content Generation 7/20/2026

he AI response executes the prompt's instructions exceptionally well. It adopts a punchy, hook-driven LinkedIn writing s...

— Thuy Hang Vo

This evaluation was conducted independently. Anthropic: Claude Sonnet 4.6 did not participate in or pay for this evaluation. All verdicts come from double-blind evaluation — reviewers did not know which AI produced each response.

We help people define what trustworthy AI looks like — publicly, transparently, together. Support this mission