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 1609 ± 1, rank #389 of 1761 rated models, from 42 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
UGI - Natural Intelligence55.79NatInt Score92.1
UGI - Writing43.85Writing Score83.9
LLM Chess (Saplin)782.5ELO82.6
Pencil Puzzle Bench - Nurimisaki13.3Direct-ask Success Rate (%)81
NonoBench46.7Overall Accuracy (%)79.8
Pencil Puzzle Bench - Slitherlink6.7Direct-ask Success Rate (%)74
IH-Benchmark87.8Overall Compliance (%)72.2
Pencil Puzzle Bench - Light Up6.7Direct-ask Success Rate (%)71
Pencil Puzzle Bench - Shikaku6.7Direct-ask Success Rate (%)71
FrontierMath - Tiers 1-326.55Accuracy (%, 290 problems)69.7
Chess Puzzles (Epoch AI)23Accuracy (%)69.6
OTIS Mock AIME 2024-2578.89Accuracy (%)62.5

Interactive version: theaggregate.ai/model?slug=gpt-5-2-low · How It Works · Data refreshed daily, snapshot 2026-09-05.