Frontier Log

Benchmark · Active

HLE-Diamond

Created by Center for AI Safety and Scale AI, 22 Sep 2026 · Higher is better · Website

020406058.8%Gemini 4 Hydrogen28.7%Inkling 2 Small020406058.8%Gemini 4 Hydrogen28.7%Inkling 2 Small

Settings and run setups

Task setToolsHarnessProviders
HLE-Diamond (1,000 questions)No toolsNot statedGoogleThinking Machines
HLE-Diamond (1,000 questions)Not statedNot statedMoonshot AI

Download image

Results

ModelScoreSource
Gemini 4 Hydrogen12 Nov 202658.8%Introducing Gemini 4 HydrogenCorroborated The same number is printed in 2 other lab documents: 1 2Run by Artificial Analysis · highest thinking · Setting from Artificial Analysis’s methodologySource image
Inkling 2 Small5 Nov 202628.7%Introducing Inkling 2 SmallCorroborated The same number is printed in 1 other lab document: 1Run by Artificial Analysis · effort 0.99 · 1M-token trajectory limit · Setting from Artificial Analysis’s methodologyOur capture

All benchmarks