AbstentionBench - underspecified context - MediQ - F1: leaderboard
Metric: Abstention F1 (%). Source: github.com. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Mistral 7B Instruct (v0.3) | 57.82 |
| 2 | Qwen 2.5 32B Instruct | 43.25 |
| 3 | Gemini 1.5 Pro | 40.32 |
| 4 | GPT-4o | 37.63 |
| 5 | Llama 3.1 405B Instruct | 35.83 |
| 6 | Llama 3.3 70B Instruct | 32.27 |
| 7 | Llama 3.1 70B Instruct | 30.51 |
| 8 | Llama 3.1 8B Instruct | 23.47 |
| 9 | Llama 3.1 8B | 22.96 |
| 10 | O1 | 17.35 |
| 11 | Llama-3.1-Tulu-3-70B-DPO | 17.22 |
| 12 | Llama 3.1 70B | 14.35 |
| 13 | Llama-3.1-Tulu-3-70B | 14.11 |
| 14 | DeepSeek R1 Distill Llama 70B | 6.98 |
| 15 | Llama-3.1-Tulu-3-8B | 5.66 |
Interactive version: theaggregate.ai/benchmark?slug=abstentionbench-underspecified-context-mediq-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.