EComStage - Query Rewrite: leaderboard
Evaluates text-based e-commerce assistance across seven perception, planning and action tasks. The overall score and task results share the main-table evaluation protocol. This series reports Query Rewrite using Task Score (%). Source: the original paper result table.
Metric: Task Score (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3 | 82.49 | #329 |
| 2 | Claude Sonnet 4 | 81.81 | #207 |
| 3 | Gemini 2.5 Pro | 81.42 | #159 |
| 4 | Qwen 3 235B A22B Instruct | 81.04 | #336 |
| 5 | InternVL3-14B | 80.82 | #506 |
| 6 | GPT-4o | 80.7 | #357 |
| 7 | glm-4-9B | 80.39 | #865 |
| 8 | Qwen 2.5 72B Instruct | 80.28 | #469 |
| 9 | Llama 3.3 70B | 80.26 | #578 |
| 10 | Qwen 3 4B Instruct | 79.97 | #661 |
| 11 | Qwen 2.5 32B Instruct | 79.74 | #509 |
| 12 | DeepSeek R1 | 79.52 | #262 |
| 13 | Qwen 3 32B | 79.37 | #449 |
| 14 | Qwen 2.5 14B Instruct | 79.06 | #654 |
| 15 | Qwen 3 235B A22B | 78.83 | #323 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ecomstage-query-rewrite · How It Works · Data refreshed daily, snapshot 2026-11-05.