Frontier Log

Benchmark · Active

GraphWalks

Created by OpenAI, 12 Apr 2025 · Higher is better · Website

020406080100Apr 2026Jul 202697.5%Mythos 558.5%GPT-5.583.3%Opus 4.856.6%Opus 4.795.5%Mythos Preview44.4%GPT-5.448.6%Opus 4.6020406080100Apr 2026Jul 202697.5%Mythos 558.5%GPT-5.583.3%Opus 4.856.6%Opus 4.795.5%Mythos Preview

Settings and run setups

DomainTask setProviders
BFSUp to 128KOpenAIGoogle
BFS256K + 1M problemsOpenAIAnthropicGoogle
BFS256K problems onlyOpenAIAnthropic
BFS1M problems onlyOpenAIAnthropic
ParentsUp to 128KOpenAI
Parents256K + 1M problemsOpenAIAnthropic
Parents256K problems onlyOpenAIAnthropic
Parents1M problems onlyOpenAIAnthropic

Download image

Results

ModelScoreSource
Claude Mythos 59 Jun 202697.5%Claude Fable 5 and Claude Mythos 5 System CardChecked A person checked this number against its source on 10 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.5 trials · run past the API's 1M-token limit
Claude Mythos Preview7 Apr 202695.5%Claude Fable 5 and Claude Mythos 5 System CardChecked A person checked this number against its source on 10 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.5 trials · run past the API's 1M-token limit · reported later
Claude Opus 4.828 May 202683.3%Claude Opus 4.8 System CardCorroborated The same number is printed in 1 other lab document: 1max effort · 5 trials · run past the API's 1M-token limit
GPT-5.523 Apr 202658.5%Introducing GPT-5.5Corroborated The same number is printed in 2 other lab documents: 1 2Our capture
Claude Opus 4.716 Apr 202656.6%Claude Opus 4.8 System CardChecked A person checked this number against its source on 10 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.5 trials · run past the API's 1M-token limit · reported later
Claude Opus 4.65 Feb 202648.6%Claude Opus 4.8 System CardChecked A person checked this number against its source on 10 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.5 trials · run past the API's 1M-token limit · reported later
GPT-5.45 Mar 202644.4%Introducing GPT-5.5Checked A person checked this number against its source on 11 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.reported laterOur capture

All benchmarks