EComStage - RAG-QA: 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 RAG-QA using Task Score (%). Source: the original paper result table.
Metric: Task Score (%). Source: arxiv.org. Saturation forecast: Around 2029. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4 | 72.13 | #207 |
| 2 | DeepSeek R1 | 71.65 | #262 |
| 3 | GPT-OSS-120B | 70.04 | #346 |
| 4 | Gemini 2.5 Pro | 69.97 | #159 |
| 5 | GLM-4 32B | 69.87 | #472 |
| 6 | Qwen 3 235B A22B Instruct | 69.76 | #336 |
| 7 | Qwen 3 235B A22B | 69.41 | #323 |
| 8 | DeepSeek V3 | 69.17 | #329 |
| 9 | Qwen 3 32B | 68.73 | #449 |
| 10 | Qwen 3 8B | 68.48 | #678 |
| 11 | Qwen 3 4B Instruct | 68.45 | #661 |
| 12 | GPT-4o | 68.24 | #357 |
| 13 | Qwen 3 30B A3B | 68.21 | #496 |
| 14 | Qwen 3 14B | 68.04 | #542 |
| 15 | Qwen 3 30B A3B Instruct | 67.75 | #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-rag-qa · How It Works · Data refreshed daily, snapshot 2026-11-05.