AbstentionBench - stale - FreshQA - Recall: leaderboard
Metric: Abstention Recall (%). Source: github.com. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama-3.1-Tulu-3-70B | 89.36 |
| 2 | Llama-3.1-Tulu-3-8B | 87.23 |
| 3 | Llama-3.1-Tulu-3-8B-DPO | 85.11 |
| 4 | Llama-3.1-Tulu-3-70B-DPO | 85.11 |
| 5 | Qwen 2.5 32B Instruct | 78.72 |
| 6 | Llama 3.1 8B Instruct | 72.34 |
| 7 | Llama 3.1 405B Instruct | 72.34 |
| 8 | Llama 3.3 70B Instruct | 63.83 |
| 9 | Llama 3.1 70B Instruct | 59.57 |
| 10 | Mistral 7B Instruct (v0.3) | 59.57 |
| 11 | Llama 3.1 8B | 53.19 |
| 12 | Llama-3.1-Tulu-3-8B-SFT | 53.19 |
| 13 | GPT-4o | 48.94 |
| 14 | Llama 3.1 70B | 42.55 |
| 15 | Gemini 1.5 Pro | 36.17 |
Interactive version: theaggregate.ai/benchmark?slug=abstentionbench-stale-freshqa-recall · How It Works · Data refreshed daily, snapshot 2026-09-19.