Frontier Log

Benchmark · Active

GraphWalks

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

020406080100Jul 2024Jan 2025Jul 2025Jan 2026Jul 202699.7%Gemini 4 Argon94.0%GPT-5.278.3%GPT-577.3%OpenAI o362.0%OpenAI o141.7%GPT-4o020406080100Jul 2024Jan 2025Jul 2025Jan 2026Jul 202699.7%Gemini 4 Argon94.0%GPT-5.278.3%GPT-572.3%GPT-4.541.7%GPT-4o

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
Gemini 4 Argon30 Sep 202699.7%Gemini 4 Argon: our next era of frontier intelligenceChecked 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.650 problems up to 128K tokens, per the model evaluation documentSource image
GPT-5.211 Dec 202594.0%Introducing GPT-5.2Corroborated The same number is printed in 1 other lab document: 1ThinkingRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-5.45 Mar 202693.0%Introducing GPT-5.4Checked 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.Our capture
GPT-5.45 Mar 202693.1%Introducing GPT-5.4 mini and nanoChecked 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.xhigh effort · reported later · another report, not chartedOur capture
GPT-57 Aug 202578.3%Introducing GPT-5 for developersChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
OpenAI o316 Apr 202577.3%Introducing GPT-5 for developersChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-5.112 Nov 202576.8%Introducing GPT-5.2Checked 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.Thinking · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-5.4 mini17 Mar 202676.3%Introducing GPT-5.4 mini and nanoChecked 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.xhigh effortOur capture
GPT-5.4 nano17 Mar 202673.4%Introducing GPT-5.4 mini and nanoChecked 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.xhigh effortOur capture
GPT-5 mini7 Aug 202573.4%Introducing GPT-5 for developersCorroborated The same number is printed in 1 other lab document: 1high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4.527 Feb 202572.3%Introducing GPT-4.1 in the APIChecked 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 laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-5 nano7 Aug 202564.0%Introducing GPT-5 for developersChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
o4-mini16 Apr 202562.3%Introducing GPT-5 for developersChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
OpenAI o117 Dec 202462.0%Introducing GPT-4.1 in the APIChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4.114 Apr 202561.7%Introducing GPT-4.1 in the APICorroborated The same number is printed in 1 other lab document: 1Run on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4.1 mini14 Apr 202561.7%Introducing GPT-4.1 in the APICorroborated The same number is printed in 1 other lab document: 1Run on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
o3-mini31 Jan 202551.0%Introducing GPT-4.1 in the APIChecked 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.high reasoning effort · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4o13 May 202441.7%Introducing GPT-4.1 in the APIChecked 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.2024-11-20 snapshot · reported laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4o mini18 Jul 202429.0%Introducing GPT-4.1 in the APIChecked 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 laterRun on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture
GPT-4.1 nano14 Apr 202525.0%Introducing GPT-4.1 in the APICorroborated The same number is printed in 1 other lab document: 1Run on the set as first published, before the creator reworded the prompts and corrected 24 parents answers on Feb 27, 2026Our capture

All benchmarks