Character Relationship Network Updating on BookWorld SNP Synthesized scenario
95.8Win RateGemini-2.5-Pro
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.5-ProBaseline=vs. BookWorld (SNP)2026.04 | 95.8 | 4.81 | |
| DeepSeek-V3.2-ThinkBaseline=vs. BookWorld (SNP)2026.04 | 94.2 | 4.69 | |
| DeepSeek-V3.2Baseline=vs. BookWorld (SNP)2026.04 | 87.9 | 4.27 | |
| Llama3.3-70BBaseline=vs. BookWorld (SNP)2026.04 | 78.3 | 3 | |
| Qwen3-32BBaseline=vs. BookWorld (SNP)2026.04 | 68.3 | 3.17 |