GPT-5.2 (Low) — benchmark results
GPT-5.2 evaluated at the low reasoning-effort setting. Provider: OpenAI. Released 2025-12-11. Access: API.
Unified ELO 1755 ± 27, rank #187 of 1776 rated models, from 41 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Natural Intelligence | 55.79 | NatInt Score | 92.8 |
| LLM Chess (Saplin) | 782.5 | ELO | 90 |
| UGI - Writing | 43.85 | Writing Score | 84.5 |
| Epoch AI - ECI | 153.72 | ECI Score | 82 |
| Pencil Puzzle Bench - Nurimisaki | 13.3 | Direct-ask Success Rate (%) | 81 |
| NonoBench | 46.7 | Overall Accuracy (%) | 79.8 |
| Pencil Puzzle Bench - Slitherlink | 6.7 | Direct-ask Success Rate (%) | 74 |
| Pencil Puzzle Bench - Light Up | 6.7 | Direct-ask Success Rate (%) | 71 |
| Pencil Puzzle Bench - Shikaku | 6.7 | Direct-ask Success Rate (%) | 71 |
| FrontierMath - Tiers 1-3 | 26.55 | Accuracy (%, 290 problems) | 69.7 |
| OTIS Mock AIME 2024-25 | 78.89 | Accuracy (%) | 62.8 |
| Vectara Hallucination Leaderboard | 91.6 | Factual Consistency Rate (%) | 60.6 |
Interactive version: theaggregate.ai/model?slug=gpt-5-2-low · How the rankings work · Data refreshed daily, snapshot 2026-07-22.