HELM Long Context - RULER SQuAD: leaderboard
HELM Long Context run of RULER SQuAD, extractive question answering embedded in long contexts (string match).
Metric: RULER String Match. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 (2025-04-14) | 88 | #250 |
| 2 | Gemini 2.0 Flash | 85 | #342 |
| 3 | GPT-4.1 Mini | 82 | #355 |
| 4 | Llama 4 Maverick Instruct FP8 | 78 | #414 |
| 5 | Gemini 2.0 Flash Lite | 77 | #450 |
| 6 | Nova Premier | 71 | #437 |
| 7 | Llama 4 Scout Instruct | 68 | #539 |
| 8 | Nova Pro | 59 | #569 |
| 9 | Nova Lite | 59 | #754 |
| 10 | GPT-4.1 Nano | 56 | #731 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-long-context-ruler-squad · How It Works · Data refreshed daily, snapshot 2026-10-20.