Text-to-JSON on STAGE-Eval
70.39PFRQwen3-4B-Thinking-2507
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Qwen3-4B-Thinking-2507Base Model=Qwen3-4B-Thinking-25072026.06 | 70.39 | 14.45 | 19.04 | 77.4 | 17.13 | |
| Qwen3-4BBase Model=Qwen3-4B2026.06 | 39.95 | 31.37 | 56.25 | 41.4 | 45.46 | |
| Llama-3.2-1B-InstructBase Model=Llama-3.2-1B-Instruct2026.06 | 12.46 | 4.07 | 21.58 | 42.18 | 19.54 | |
| Qwen3-4B + GlaiveBase Model=Qwen3-4B, Training Data=Glaive2026.06 | 11.01 | 32.43 | 74.07 | 18 | 58.71 | |
| Qwen2.5-3BBase Model=Qwen2.5-3B2026.06 | 10.03 | 21.5 | 56.56 | 25.02 | 45.87 | |
| Qwen3-4B + ScrapeGraphAIBase Model=Qwen3-4B, Training Data=ScrapeGraphAI2026.06 | 5.64 | 35.53 | 83.2 | 10.29 | 63.24 | |
| Llama-3.2-3B-InstructBase Model=Llama-3.2-3B-Instruct2026.06 | 4.66 | 11.32 | 35.72 | 27.03 | 36.05 | |
| Qwen3-4B + JSONSchemaBenchBase Model=Qwen3-4B, Training Data=JSONSchemaBench2026.06 | 1.69 | 31.73 | 89.74 | 3.87 | 48.54 | |
| Llama-3.2-1B-Instruct + STAGEBase Model=Llama-3.2-1B-Instruct, Training Data=STAGE2026.06 | 1.18 | 53.47 | 93.42 | 4.38 | 80.72 | |
| Llama-3.2-3B-Instruct + STAGEBase Model=Llama-3.2-3B-Instruct, Training Data=STAGE2026.06 | 1.18 | 59.69 | 95.02 | 3.53 | 82.72 | |
| Qwen2.5-3B + STAGEBase Model=Qwen2.5-3B, Training Data=STAGE2026.06 | 0.47 | 60.87 | 96.36 | 2.5 | 84.95 | |
| Qwen3-4B SFT + STAGEBase Model=Qwen3-4B SFT, Training Data=STAGE2026.06 | 0.35 | 74.27 | 98.24 | 1.29 | 90.69 |