Skip to content
Evaluation sources

FrontierSWE v2

Proximal Labs / Coding · Long-running engineering and research tasks

Official evaluation
Across NexGateHub AI News Observing
Evidence budgetNot scored
Upstream data timeUnverified
Last successful syncCollection has not started

What it measures and how

Only the common v2 Proximus harness is compared. Earlier versions and best-of runs under different conditions remain separate.

How this evidence is used

Under review: long-task coverage of new models is limited. Stable raw-data access and run configurations are still needed.

Evaluation limits & attribution

Under review: long-task coverage of new models is limited. Stable raw-data access and run configurations are still needed.

Data license: Leaderboard-data usage terms need confirmation.

Results are published by Proximal Labs. Original scores and the NexGateHub consensus index use different scales and cannot be added directly.