Frontier Log

OpenAI

GPT-5.6 Terra

Announced on openai.com

Benchmark results

BenchmarkScoreSource check
SWE-bench ProThe source doesn't say whether this is the public or the commercial set63.4%
CheckedA person checked this number against its source on 26 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
GPQA Diamond92.9%
CheckedA person checked this number against its source on 26 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
BrowseComp87.5%
CheckedA person checked this number against its source on 26 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
AutomationBench15.2%
CheckedA person checked this number against its source on 26 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
OSWorld 2.0Offline set, partial score · set, scoring and task release not stated50.2%
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
DeepSWE v1.169.6%
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
Terminal-Bench 2.187.4%
CorroboratedThe same number is printed in 3 other lab documents: 1 2 3
ARC-AGI-30.8%
CheckedA person checked this number against its source on 26 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced.
Sources

Reported by OpenAI

Related events

All models