Introducing Claude Opus 4.8
Open paper
Benchmark
Model
Rank
Results
code-generation-on-swe-bench-pro
Claude Opus 4.8
–
Resolved: 69.2
code-generation-on-terminal-bench-2-1
Terminus-2 + Claude Opus 4.8
–
Accuracy: 74.6
Rank counts only results with a code link.