Gemini 1.5 Pro (Sept) — benchmark results

September 2024 Gemini 1.5 Pro snapshot, tracked when sources report the dated API model. Provider: Google. Released 2024-09-24. Access: API.

Unified ELO 1552 ± 20, rank #616 of 1776 rated models, from 22 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
RewardBench86.78Score (%)86.1
Deception Effectiveness (Lechmazur)0.94Deception Score76.5
RewardBench Chat Hard76.97Accuracy (%)76.1
LLM Public Goods Game24.64Avg. Contribution (%)73.7
RewardBench Reasoning90.22Accuracy (%)73.4
RewardBench Safety85.81Accuracy (%)66.7
Confabulation Leaderboard (Lechmazur)16.83Confabulation rate % (lower is better)61.9
AA MATH-50087.6Accuracy (%)60.4
RewardBench Chat94.13Accuracy (%)48.9
AA MMLU-Pro75.02Accuracy (%)48.3
Deception Resistance (Lechmazur)0.62Vulnerability Score (lower is better)47.1
NYT Connections Original22.7Score (%)43.3

Interactive version: theaggregate.ai/model?slug=gemini-1-5-pro-sept · How the rankings work · Data refreshed daily, snapshot 2026-07-22.