Frontier Log

Benchmark · InactiveSuperseded by OSWorld 2.1

OSWorld 2.0 (v2026.08.08)

Created by XLANG Lab, 8 Aug 2026 · Higher is better · Website

020406080Jul 2026Oct 202672.6%GPT-6 Astra69.2%Gemini 4 Argon60.5%GPT-6 Sol65.7%GPT-5.6 Sol020406080Jul 2026Oct 202672.6%GPT-6 Astra69.2%Gemini 4 Argon71.4%GPT-6.1 Sol60.5%GPT-6 Sol65.7%GPT-5.6 Sol

Results

ModelScoreSource
GPT-6 Astra3 Sep 2026 · OpenAI72.6%GPT-6 Astra: A new generation of intelligenceChecked A person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.Our capture
GPT-6 Astra3 Sep 2026 · OpenAI62.2%Introducing GPT-6.1 SolCorroborated The same number is printed in 1 other lab document: 1low effort; reported later; another report, not chartedOur capture
GPT-6 Astra3 Sep 2026 · OpenAI69.3%Introducing GPT-6.1 SolCorroborated The same number is printed in 1 other lab document: 1medium effort; reported later; another report, not chartedOur capture
GPT-6 Astra3 Sep 2026 · OpenAI70.0%Introducing GPT-6.1 SolCorroborated The same number is printed in 1 other lab document: 1high effort; reported later; another report, not chartedOur capture
GPT-6 Astra3 Sep 2026 · OpenAI71.3%Introducing GPT-6.1 SolCorroborated The same number is printed in 1 other lab document: 1xhigh effort; reported later; another report, not chartedOur capture
GPT-6 Astra3 Sep 2026 · OpenAI73.5%Introducing GPT-6.1 SolCorroborated The same number is printed in 1 other lab document: 1max effort; reported later; another report, not chartedOur capture
GPT-6.1 Sol29 Sep 2026 · OpenAI71.4%Introducing GPT-6.1 SolChecked A person checked this number against its source on 8 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.max effortOur capture
GPT-6.1 Sol29 Sep 2026 · OpenAI59.0%Introducing GPT-6.1 SolChecked A person checked this number against its source on 8 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.low effort; another report, not chartedOur capture
GPT-6.1 Sol29 Sep 2026 · OpenAI66.8%Introducing GPT-6.1 SolChecked A person checked this number against its source on 8 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.medium effort; another report, not chartedOur capture
GPT-6.1 Sol29 Sep 2026 · OpenAI69.6%Introducing GPT-6.1 SolChecked A person checked this number against its source on 8 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.high effort; another report, not chartedOur capture
GPT-6.1 Sol29 Sep 2026 · OpenAI69.4%Introducing GPT-6.1 SolChecked A person checked this number against its source on 8 Nov 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.xhigh effort; another report, not chartedOur capture
Gemini 4 Argon30 Sep 2026 · Google69.2%Gemini 4 Argon: our next era of frontier intelligenceChecked A person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.Source image
GPT-5.6 Sol9 Jul 2026 · OpenAI65.7%GPT-6 Astra: A new generation of intelligenceChecked A person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.reported laterOur capture
GPT-6 Sol22 Sep 2026 · OpenAI60.5%Introducing GPT-6 Sol and LunaChecked A person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published. It doesn't confirm that the score was reproduced.xhigh effort

All benchmarks