GPT-5.6 Luna: benchmark results

OpenAI's fast, low-cost GPT-5.6 tier. Provider: OpenAI. Released 2026-07-09. Access: API.

Unified ELO 1690 ± 1, rank #67 of 1392 rated models, from 210 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AgentBattler56.21Pooled Score (%)100
Coarena - Task Completion89.1Completion Rate (%)100
YapBench27.8YapIndex (lower is better)98.1
Vals AI CyberBench83.9Accuracy (%)95.7
Creative Writing v31932.2Elo score (self-reported)95.5
Clerk LLM Leaderboard83.6Avg score (%)95.2
BenchmarkList ECI146.95Capability Index (ECI)94
AIIQ Composite IQ129Composite IQ (self-reported)93.5
Vellum - HumanEval93Pass@1 (%)93.5
ZeroEval GPQA Diamond92.3GPQA Diamond Score93.5
Nejumi 4 - GLP - Syntactic Analysis87.83Score (%)92.2
LLM Stats Score45.38LLM Stats Score (conservative rating)92

Interactive version: theaggregate.ai/model?slug=gpt-5-6-luna · How It Works · Data refreshed daily, snapshot 2026-09-05.