E-commerce Last Exam - Travel: leaderboard
E-commerce Last Exam restricted to travel episodes: itinerary search and booking against a simulated provider, scored by mean reward over the episode.
Metric: Mean Reward (%). Source: huggingface.co. Saturation forecast: Around 2033. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K3 | 48.11 | #30 |
| 2 | Grok 4.5 | 46.73 | #48 |
| 3 | Claude Opus 5 | 45.05 | #12 |
| 4 | Claude Opus 4.7 | 44.82 | #49 |
| 5 | Claude Opus 4.8 | 44.72 | #37 |
| 6 | Gemini 3.7 Flash | 44.51 | #23 |
| 7 | GLM-5.2 | 44.16 | #73 |
| 8 | GPT-5.5 | 43.9 | #31 |
| 9 | Gemini 3.6 Flash | 43.78 | #51 |
| 10 | Claude Sonnet 5 | 43.68 | #57 |
| 11 | GPT-5.6 Sol | 43.61 | #19 |
| 12 | Qwen 3.8 Max | 43.37 | #35 |
| 13 | MiMo-V2.5-Pro | 43.37 | #158 |
| 14 | Qwen 3.8 Flash | 43.36 | #50 |
| 15 | Claude Opus 4.6 | 43.34 | #63 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=e-commerce-last-exam-travel · How It Works · Data refreshed daily, snapshot 2026-10-20.