Text-to-Stage on PDNC 2 (val)
96Quote AttributionClaude Sonnet 3.5
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Claude Sonnet 3.5Model Category=Proprietary models2026.03 | 96 | 97 | 87 | 85 | 77 | 51 | 82 | |
| Claude Sonnet 4Model Category=Proprietary models2026.03 | 96 | 97 | 85 | 83 | 69 | 50 | 80 | |
| Llama-4-MaverickModel Category=Open-source models2026.03 | 96 | 99.9 | 74.3 | 85.6 | 87.4 | 55.7 | 83.1 | |
| GPT-4oModel Category=Proprietary models2026.03 | 95 | 99 | 68 | 83 | 84 | 51 | 80 | |
| GPT-4.1Model Category=Proprietary models2026.03 | 94 | 99 | 65 | 84 | 90 | 51 | 81 | |
| Claude Sonnet 3.7Model Category=Proprietary models2026.03 | 94 | 97 | 88 | 83 | 64 | 49 | 79 | |
| Gemini-2.0-flashModel Category=Proprietary models2026.03 | 93 | 96 | 68 | 81 | 86 | 51 | 79 | |
| Spatializer-SFTModel Category=Our models2026.03 | 92.8 | 98.2 | 77.7 | 87.2 | 96.1 | 57 | 84.8 | |
| Llama-3.3-70B-itModel Category=Open-source models2026.03 | 92.5 | 98.9 | 71.6 | 86.9 | 90.2 | 60.4 | 83.4 | |
| Llama-4-ScoutModel Category=Open-source models2026.03 | 92 | 99 | 71.1 | 86.8 | 90.5 | 61.4 | 83.4 | |
| Spatializer-GRPOModel Category=Our models2026.03 | 91.8 | 99.1 | 92.1 | 92.5 | 98.2 | 57.4 | 88.5 | |
| Gemma-3-27b-itModel Category=Open-source models2026.03 | 89.1 | 98.1 | 80.3 | 73.7 | 78.3 | 60.4 | 80 | |
| Llama-3.3-8B-itModel Category=Open-source models2026.03 | 86.8 | 96.8 | 82.2 | 84.3 | 83.7 | 56.2 | 81.7 | |
| Qwen3-8bModel Category=Open-source models2026.03 | 83.8 | 95.1 | 68.1 | 79.7 | 91.2 | 51.1 | 78.2 | |
| Llama-3.1-8B-itModel Category=Open-source models2026.03 | 27 | 31.6 | 23 | 27.6 | 29.3 | 15.1 | 25.6 |