Datacurve
11 posts

Datacurve
@datacurve
Research and data to advance frontier models.

Opus 4.8 is now on DeepSWE. On the default high thinking effort, it scores 6% higher than Opus 4.7 xhigh, while also lowering average cost per task.


Today we’re releasing DeepSWE, a new standard for agentic coding benchmarks. On public leaderboards, top models often look relatively close in capability. DeepSWE shows where they actually diverge, reflecting the realistic experience of developers in their day-to-day work.


Today we’re releasing DeepSWE, a new standard for agentic coding benchmarks. On public leaderboards, top models often look relatively close in capability. DeepSWE shows where they actually diverge, reflecting the realistic experience of developers in their day-to-day work.







