Coding / Agent
FreshHigh confidenceSWE-benchSWE-bench bash-only · result 2025-12-11Updated: 8/10/2026
71.8% resolved
#10 · 80th percentile
OpenAI·GPT
GPT-5.2 High
5.2 High
Aliases: gpt-5.2-high, gpt-5-2-high
Canonical ID
openai-gpt-5.2-highIndependent sources
1
Composite eligibility
Needs more sources
Public benchmark scores
Source ranks stay intact; incompatible raw scores are not added.
Why it ranks here
Only 1 independent source is verified, so the page keeps the source rank and does not invent a composite total.
Historical movement
Historical snapshots are available on the source score cards.