HSCodeComp - 10-digit: leaderboard
HS customs code prediction scored at full 10-digit statistical-suffix precision — the complete code a customs declaration requires, and the hardest level, where weaker models score near zero.
Metric: Exact Match Accuracy (%). Source: arxiv.org. Saturation forecast: Around January 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 28.96 | #98 |
| 2 | Gemini 2.5 Pro | 23.42 | #152 |
| 3 | GPT-4o | 18.35 | #343 |
| 4 | Kimi K2 | 12.18 | #243 |
| 5 | Claude Sonnet 4 | 11.87 | #199 |
| 6 | DeepSeek R1 | 6.65 | #251 |
| 7 | DeepSeek V3 | 6.49 | #321 |
| 8 | Qwen Max | 3.8 | #382 |
| 9 | Qwen 3 235B A22B | 1.74 | #313 |
| 10 | O3 Mini | 1.27 | #272 |
| 11 | Qwen 3 32B | 0.32 | #443 |
| 12 | QwQ-32B | 0.16 | #421 |
| 13 | Qwen 2.5 72B | 0.16 | #499 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=hscodecomp-10-digit · How It Works · Data refreshed daily, snapshot 2026-10-20.