ProEvolve E-Commerce (Completion): leaderboard
Metric: State-wise task success rate (0-1, times 100) on the versions evolved by capability addition (completion), ProEvolve's e-commerce tool-calling domain (WebShop products, 50 synthetic users): 50 version-linked evolution trajectories with task sandboxes of balanced easy, medium and hard difficulty generated from each version's environment graph; the agent serves a Claude-Opus-4.5 customer simulator that follows state-wise instructions, without memory of earlier versions (baseline); higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 65 | #91 |
| 2 | Qwen 3 235B A22B 2507 (Thinking) | 63 | #253 (Qwen 3 235B A22B 2507) |
| 3 | Gemini 2.5 Pro | 57 | #145 |
| 4 | DeepSeek V3.2 | 52 | #198 |
| 5 | Claude Opus 4.5 (20251101) | 46 | #70 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=proevolve-e-commerce-completion · How It Works · Data refreshed daily, snapshot 2026-10-11.