UnpredictaBench - Text: leaderboard
Metric: KS@100 (%): share of the 160 Text problems (named or described canonical distributions) for which 100 samples from the model (100 independent calls at temperature 1.0, reasoning effort none, up to 5 retries for a parseable answer) are not rejected by a two-sample Kolmogorov-Smirnov test against ground-truth samples at p < 0.0001; higher is better. Source: arxiv.org. Saturation forecast: Around 2039. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3.2 (Non-reasoning) | 28.13 |
| 2 | GPT-5.4 (Non-reasoning) | 25.63 |
| 3 | Llama 3.1 70B Instruct | 23.13 |
| 4 | GPT-4o | 21.25 |
| 5 | Ministral-3-3B-Instruct-2512 | 21.25 |
| 6 | Nemotron 3 Nano 30B A3B (Non-reasoning) | 20.63 |
| 7 | Qwen 3.5 2B (Non-reasoning) | 14.38 |
| 8 | Qwen 3 32B (Non-reasoning) | 11.88 |
| 9 | Llama 3.2 1B Instruct | 10 |
| 10 | GPT-4o Mini | 9.38 |
| 11 | OLMo 3 7B Instruct | 7.5 |
| 12 | Grok 4.1 Fast (Non-reasoning) | 6.88 |
| 13 | Claude Sonnet 4.6 | 5.03 |
| 14 | Qwen 3.5 397B A17B (Non-reasoning) | 3.8 |
| 15 | Llama 3.1 8B Instruct | 3.75 |
Interactive version: theaggregate.ai/benchmark?slug=unpredictabench-text · How It Works · Data refreshed daily, snapshot 2026-09-29.