HELM Classic - MS MARCO TREC: leaderboard
The "MS MARCO TREC" subtask of HELM Classic. A scenario from Stanford HELM Classic, the holistic evaluation suite scoring language models across accuracy, robustness and calibration on a broad set of tasks.
Metric: NDCG@10 (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | text-davinci-002 | 66.37 | #1225 |
| 2 | text-davinci-003 | 64.38 | #1146 |
| 3 | text-curie-001 | 50.72 | #1632 |
| 4 | OPT-66B | 48.2 | #1603 |
| 5 | text-babbage-001 | 44.92 | #1648 |
| 6 | OPT-175B | 44.81 | #1538 |
| 7 | gpt-neox-20B | 39.82 | #1600 |
| 8 | davinci | 37.82 | #1553 |
| 9 | gpt-j-6B | 34.5 | #1609 |
| 10 | text-ada-001 | 30.16 | #1655 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-classic-ms-marco-trec · How It Works · Data refreshed daily, snapshot 2026-10-20.