Model

Inkling

Thinking Machines

Knowledge & Reasoning

Knowledge & Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
CritPt5.4accuracyN/AIndependentlowepoch.ai
GPQA Diamond88.3accuracyN/AIndependenthighepoch.ai
SimpleQA Verified40.3accuracyN/AIndependenthighepoch.ai

Coding

Coding
BenchmarkScoreMetricHarnessSourceConfidenceLink
SciCode47.0scoreN/AIndependentmediumartificialanalysis.ai

Multimodal

Multimodal
BenchmarkScoreMetricHarnessSourceConfidenceLink
PlotGround-1k32.0accuracyN/AIndependentmediumarxiv.org

Safety & Alignment

Safety & Alignment
BenchmarkScoreMetricHarnessSourceConfidenceLink
DistressBench0.8scoreN/AIndependentmediumarxiv.org
RT-Safe0.0safe_success_rateN/AIndependentmediumarxiv.org

Abstract Reasoning

Abstract Reasoning
BenchmarkScoreMetricHarnessSourceConfidenceLink
ARC-AGI-236.5scoreN/AIndependentlowepoch.ai

Reproduced by the community

Loading community runs…