AbstentionBench - underspecified context - GSM8K - F1: leaderboard
Metric: Abstention F1 (%). Source: github.com. 23 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 (Low) | 97.73 |
| 2 | O1 (High) | 97.68 |
| 3 | O1 | 97.6 |
| 4 | Llama 3.1 405B Instruct | 97.34 |
| 5 | Qwen 2.5 32B Instruct | 97.16 |
| 6 | Llama 3.1 70B Instruct | 97.03 |
| 7 | Gemini 1.5 Pro | 96.99 |
| 8 | Llama 3.1 8B Instruct | 96.07 |
| 9 | GPT-4o | 95.84 |
| 10 | Llama 3.3 70B Instruct | 95.34 |
| 11 | Llama-3.1-Tulu-3-70B-DPO | 94.25 |
| 12 | Llama-3.1-Tulu-3-70B | 93.5 |
| 13 | Mistral 7B Instruct (v0.3) | 91.55 |
| 14 | Llama-3.1-Tulu-3-8B-DPO | 85.86 |
| 15 | Llama-3.1-Tulu-3-8B | 85.4 |
Interactive version: theaggregate.ai/benchmark?slug=abstentionbench-underspecified-context-gsm8k-f1 · How It Works · Data refreshed daily, snapshot 2026-09-19.