EComStage - Query Match: 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 Match 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 | Qwen 3 235B A22B | 98.81 | #323 |
| 2 | Gemini 2.5 Pro | 98.75 | #159 |
| 3 | Qwen 3 14B | 98.75 | #542 |
| 4 | Qwen 3 235B A22B Instruct | 98.75 | #336 |
| 5 | Qwen 2.5 72B Instruct | 98.7 | #469 |
| 6 | DeepSeek R1 | 98.7 | #262 |
| 7 | Qwen 2.5 32B Instruct | 98.5 | #509 |
| 8 | Qwen 3 32B | 98.44 | #449 |
| 9 | Claude Sonnet 4 | 98.44 | #207 |
| 10 | GPT-OSS-120B | 98.39 | #346 |
| 11 | DeepSeek V3 | 98.34 | #329 |
| 12 | Llama 3.3 70B | 98.08 | #578 |
| 13 | Qwen 2.5 14B Instruct | 97.98 | #654 |
| 14 | Qwen 3 30B A3B | 97.98 | #496 |
| 15 | Qwen 3 30B A3B Instruct | 97.92 | #487 |
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-match · How It Works · Data refreshed daily, snapshot 2026-11-05.