Summarization on SummaryBench Romeo and Juliet
100FRMagentic-One
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Magentic-OneLLM Backend=GPT-5.4 mini2026.05 | 100 | — | |
| MegaAgentLLM Backend=GPT-4.1 mini2026.05 | 80 | 472 | |
| DirectLLM Backend=GPT-5.4 mini2026.05 | 0 | 19 | |
| APWALLM Backend=GPT-5.4 mini2026.05 | 0 | 157 |