Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Task Generation on SWE RL-training repositories
Loading...
26.64
Utility
PROPEL
21.5856
22.8978
24.21
25.5222
Jun 10, 2026
Utility
Validity
Self-BLEU (3-gram)
Distinct-3
Top-Topic Coverage
Updated 1mo ago
Evaluation Results
Method
Method
Links
Utility
Validity
Self-BLEU (3-gram)
Distinct-3
Top-Topic Coverage
PROPEL
Solver=27B
2026.06
26.64
80.42
88.2
22.6
69.4
Base
Solver=27B
2026.06
21.78
77.69
77.8
31.2
73.7
Feedback
Search any
task
Search any
task
