Technological requirement identification on Scientific Contribution Graph 1.0 (pre-cutoff)
53MAPCLAUDE OPUS 4.6
Evaluation Results
| Method | Links | |
|---|---|---|
| CLAUDE OPUS 4.6Knowledge Cutoff=Aug 2025, Cost (per 1k)=$235, # Samples (pre/post)=1364 / 2052026.05 | 53 | |
| CLAUDE SONNET 4.5Knowledge Cutoff=Jul 2025, Cost (per 1k)=$160, # Samples (pre/post)=1337 / 2722026.05 | 45 | |
| GPT-5-MINIKnowledge Cutoff=May 2024, Cost (per 1k)=$22, # Samples (pre/post)=817 / 10902026.05 | 42 | |
| CLAUDE HAIKU 4.5Knowledge Cutoff=Jun 2025, Cost (per 1k)=$45, # Samples (pre/post)=1337 / 2722026.05 | 41 | |
| GPT-5.4Knowledge Cutoff=Aug 2025, Cost (per 1k)=$106, # Samples (pre/post)=1364 / 2052026.05 | 40 | |
| GPT-5.4-MINIKnowledge Cutoff=Aug 2025, Cost (per 1k)=$29, # Samples (pre/post)=1364 / 2052026.05 | 37 | |
| GPT-OSS-120BKnowledge Cutoff=Jun 2024, Cost (per 1k)=$20, # Samples (pre/post)=845 / 10762026.05 | 25 | |
| GPT-5-NANOKnowledge Cutoff=May 2024, Cost (per 1k)=$5, # Samples (pre/post)=817 / 10902026.05 | 25 | |
| GPT-4O-MINIKnowledge Cutoff=Oct 2023, Cost (per 1k)=$16, # Samples (pre/post)=617 / 12792026.05 | 16 |