Model

claude-opus-4-20250514-32K

Anthropic · closed weights

anthropic.com

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
Aider Polyglot72.0percent_correctN/AIndependentmediumaider.chat

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
GeoBench49.0acw_country_%N/AIndependentmediumgeobench.org

Reproduced by the community

Loading community runs…