HREF: leaderboard
HREF evaluates instruction-following models with human response-guided automatic evaluation across 11 task categories.
Metric: Average HREF Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 34 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 70B Instruct | 48.98 | #561 |
| 2 | Mistral Large 2 (Jul) | 48.39 | #484 |
| 3 | Qwen 2.5 72B Instruct | 46.21 | #454 |
| 4 | Llama-3.1-Tulu-3-70B | 43.68 | #482 |
| 5 | Mistral-Small-Instruct-2409 | 42.87 | #702 |
| 6 | Qwen 1.5 110B Chat | 40.76 | #691 |
| 7 | Llama 3.1 8B Instruct | 38.57 | #1035 |
| 8 | OLMo-2-1124-13B-Instruct | 35.6 | #1100 |
| 9 | Yi 1.5 34B Chat | 35.1 | #800 |
| 10 | Qwen 2 72B Instruct | 33.71 | #576 |
| 11 | Llama-3.1-Tulu-3-8B | 33.54 | #849 |
| 12 | Phi-3-medium-4k-instruct | 30.91 | #952 |
| 13 | OLMo-2-1124-7B-Instruct | 28.49 | #1367 |
| 14 | Llama 2 70B Chat (HF) | 23.9 | #1264 |
| 15 | tulu-2-dpo-70B | 22.67 | #1052 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=href · How It Works · Data refreshed daily, snapshot 2026-10-20.