SupChain-Bench (QA, With Context): leaderboard
Metric: Accuracy (%) on SupChain-Bench's 435 supply-chain knowledge questions (141 multiple-answer, 147 single-choice and 147 true/false items drafted from anonymized industry documents by a three-model pipeline and verified by domain experts); one prompt template, answer extracted by regular expression; the source document (first 2,000 characters) is prefixed to each question; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 83.91 | #91 |
| 2 | GPT-5 Mini | 83.91 | #176 |
| 3 | GPT-4.1 | 82.76 | #240 |
| 4 | Claude Sonnet 4 | 82.07 | #194 |
| 5 | Gemini 2.5 Pro | 80.69 | #145 |
| 6 | Qwen 3 Max | 80.69 | #201 |
| 7 | Claude Opus 4 | 80.69 | #155 |
| 8 | Kimi K2 | 80.68 | #236 |
| 9 | GPT-4.1 Mini | 80.46 | #346 |
| 10 | Claude 3.5 Sonnet | 80.45 | #337 |
| 11 | O3 Mini | 80.23 | #266 |
| 12 | DeepSeek V3 | 79.77 | #312 |
| 13 | DeepSeek R1 | 78.16 | #245 |
| 14 | Claude 3.7 Sonnet | 78.16 | #241 |
| 15 | Qwen 3 30B A3B | 77.93 | #488 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=supchain-bench-qa-with-context · How It Works · Data refreshed daily, snapshot 2026-10-11.