Frontier Log

OpenAIModel

GPT-5.4

Made by
OpenAI
Released
Announced in
Introducing GPT-5.4openai.com
Event
GPT-5.4
API record
/api/v1/models/gpt-5-4

Benchmark results

BenchmarkScoreSource check
Humanity's Last ExamNo tools39.8%
CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 5 other lab documents, the first 3 linked: 1 2 3
Humanity's Last ExamWith tools52.1%
CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 5 other lab documents, the first 3 linked: 1 2 3
SWE-bench Pro57.7%
CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 8 other lab documents, the first 3 linked: 1 2 3
GPQA Diamondxhigh effort92.8%
CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
xhigh effort · reported later, in Introducing GPT-5.4 mini and nano93.0%
CheckedA person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
BrowseComp82.7%
CheckedA person checked this number against its source on 17 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 5 other lab documents, the first 3 linked: 1 2 3
ARC-AGI-2ARC Prize Verified73.3%
CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 2 other lab documents: 1 2
Terminal-Bench 2.0harness not stated75.1%
CheckedA person checked this number against its source on 16 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 6 other lab documents, the first 3 linked: 1 2 3
OSWorld-Verified75.0%
CheckedA person checked this number against its source on 18 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 5 other lab documents, the first 3 linked: 1 2 3
ARC-AGI-1ARC Prize Verified93.7%
CheckedA person checked this number against its source on 18 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
CorroboratedThe same number is printed in 2 other lab documents: 1 2
Sources

Reported by OpenAI

Events

Stories

All models