Know2Guess - Answer Accuracy: leaderboard
Know2Guess (2026) tests knowledge-boundary behavior across five domains: models should answer questions with a unique supported answer (zones A to C) and abstain on programmatically built synthetic-unknown questions (zone D). This board is answer accuracy in percent on the answer-expected zones A to C of the test split; higher is better.
Metric: Answer accuracy (0-1 scaled to %) on the answer-expected zones A to C (questions with a unique supported answer, five domains); standard prompt on the 1,080-item test split, greedy decoding with at most 128 or 160 output tokens, no retrieval or tools; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 9B (Non-reasoning) | 48.33 | #803 (Qwen 3.5 9B) |
| 2 | Mistral Small 3.2 | 44.18 | #1591 |
| 3 | Llama 3 8B Instruct | 41.41 | #4612 |
| 4 | Qwen 3 8B (Non-reasoning) | 34.26 | #1830 (Qwen 3 8B) |
| 5 | Qwen 2.5 3B Instruct | 27.68 | #5150 |
| 6 | Qwen 2.5 1.5B Instruct | 8.19 | #7470 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=know2guess-answer-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-13.