List your company

Research Replication on ReplicationBench

17.8Pass@1MTL (I)
5.4248.63711.8515.063Apr 15, 2026

Evaluation results

MethodLinks
MTL (I)Model Backbone=DeepSeek V3.22026.0417.8
ZeroShotModel Backbone=DeepSeek V3.22026.0414.4
MTL (I)Model Backbone=Qwen3-Coder-480B-A35B-Instruct2026.0413
ZeroShotModel Backbone=Qwen3-Coder-480B-A35B-Instruct2026.0412.6
MTL (I)Model Backbone=GPT-5-mini2026.0411.9
MTL (W)Model Backbone=GPT-5-mini2026.048.1
MTL (S)Model Backbone=GPT-5-mini2026.047.8
MTL (T)Model Backbone=GPT-5-mini2026.046.3
ZeroShotModel Backbone=GPT-5-mini2026.045.9