GPT-4 Preview (0125): benchmark results
January 25, 2024 GPT-4 Turbo preview snapshot (gpt-4-0125-preview). Provider: OpenAI. Released 2024-01-25. Access: API.
Aggregate Rating 1587 ± 11, rank #486 of 1656 rated models, from 110 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BiGGen-Bench | 4.19 | Average Score (1-5) | 99 |
| LogicKor - Single-Turn | 9.33 | Score (0-10) | 98.7 |
| LogicKor - Reasoning | 9.35 | Score (0-10) | 97.7 |
| LogicKor | 9.13 | Score (0-10) | 97.3 |
| Enkrypt AI - Safety Risk | 15.97 | Risk Score | 95.6 |
| LogicKor - Grammar | 8.21 | Score (0-10) | 95 |
| SAD - Facts (Situating Prompt) | 72.61 | Score (%, higher is better) | 95 |
| MAGI-Hard | 76.83 | Accuracy (%, 2024-05 snapshot) | 94.8 |
| LogicKor - Understanding | 9.78 | Score (0-10) | 94.7 |
| LogicKor - Math | 8.21 | Score (0-10) | 94.3 |
| LogicKor - Writing | 9.57 | Score (0-10) | 93.3 |
| LogicKor - Multi-Turn | 8.92 | Score (0-10) | 92.9 |
Interactive version: theaggregate.ai/model?slug=gpt-4-preview-0125 · How It Works · Data refreshed daily, snapshot 2026-10-20.