SommBench - Food and Wine Pairing: leaderboard
Metric: Food and wine pairing: Matthews correlation coefficient (x 100, from -100 to 100) of yes/no judgments on 1,000 sommelier-labelled food and wine pairings in English, zero-shot, temperature 0, one run; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 39 | #237 |
| 2 | GPT-4.1 | 25 | #240 |
| 3 | GPT-OSS-20B (Low) | 25 | #499 (GPT-OSS-20B) |
| 4 | GPT-4.1 Mini | 20 | #346 |
| 5 | GPT-OSS-120B (Low) | 20 | #330 (GPT-OSS-120B) |
| 6 | GPT-4o | 19 | #333 |
| 7 | GPT-OSS-120B (Medium) | 18 | #330 (GPT-OSS-120B) |
| 8 | GPT-5 | 17 | #91 |
| 9 | GPT-OSS-20B (Medium) | 16 | #499 (GPT-OSS-20B) |
| 10 | Gemini 2.5 Pro | 12 | #145 |
| 11 | GPT-4o Mini | 12 | #588 |
| 12 | GPT-OSS-120B (High) | 11 | #330 (GPT-OSS-120B) |
| 13 | Gemini 2.5 Flash Lite | 6 | #413 |
| 14 | Grok 4 Fast | 5 | #242 |
| 15 | Grok 4 | 1 | #169 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=sommbench-food-and-wine-pairing · How It Works · Data refreshed daily, snapshot 2026-10-11.