Expert-SWE (Internal)

OpenAI internal frontier software engineering evaluation for long-horizon coding tasks with a median estimated human completion time of 20 hours.

2rows
scoreprimary metric
2026-04-23sampled

Metadata

ID
openai_expert_swe_internal
Category
Coding
Release
Unknown

Metrics

Score

Latest Results

Provider-published launch-post benchmark scores parsed from OpenAI's evaluation tables. Rows are marked self-reported and should be interpreted as source claims unless independently reproduced. OpenAI notes GPT evals were run with reasoning effort set to xhigh in a research environment.

Rank Subject Score Model Match Provenance Sampled
1 GPT-5.5 73.1% GPT-5.5
openai-gpt-5.5
 		Launch post 2026-04-23
2 GPT-5.4 68.5% GPT-5.4
openai-gpt-5.4
 		Launch post 2026-04-23