Expert-SWE (OpenAI internal)

OpenAI internal long-horizon coding eval with ~20h median human completion time.

The best 2 · 2 models tested

Top 2 · best setting per model · 2 models, 2 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported

Expert-SWE (OpenAI internal) leaderboard
				Thinking levelSetting
1	GPT-5.5	OpenAI	73.1%	Extra highreasoning effort=xhigh	Maker's own figureVendor-reportedOpenAI ↗	—	23 Apr 2026
2	GPT-5.4	OpenAI	68.5%	Extra highreasoning effort=xhigh	Maker's own figureVendor-reportedOpenAI ↗	—	23 Apr 2026