LLM Safety and Ethics Index

← All models

Developer

Google

52 indexed models; 29 currently meet the evidence threshold for the overall ranking. Together they have results from 70 evaluations.

Official Google model information ↗

Company governance evidence

Google is represented at +0.74 SD relative to the matched Future of Life Institute edition. Provenance: Direct. This company-level evidence contributes 10% of overall model rank.

Models by Google

ModelEvalsComponentsRankRelease date
Gemini 3.1 Pro Preview207/7182026-02-19
Gemini 3.6 Flash55/7192026-07-21
Gemini 2.5 Pro Exp43/7352025-03-25
Gemini 1.0 Pro86/7572023-12-06
Gemini 3.5 Flash207/7662026-05-19
Gemini 1.5 Pro207/7702024-02-15
Gemini 3 Pro Preview227/7832025-11-18
Gemini 3.5 Flash Lite64/7912026-07-21
Gemini 1.5 Flash156/7932024-05-14
Gemini 2.5 Pro337/71042025-03-25
Gemma 4 31B It87/71152026-03-11
Gemma 4 26B A4B It55/71162026-03-11
Gemini 2.5 Flash317/71212025-04-17
Gemini 3.1 Flash Lite207/71252026-03-03
Gemini 2 Flash75/71302024-12-11
Gemini 3 Flash Preview156/71462025-12-17
Gemini 2 Pro Preview63/71532025-02-05
Gemma 2B It56/71582024-02-21
Gemini 2.0 Flash126/71622024-12-11
Gemma 7B It77/71672024-02-21
Gemma 2 9B It77/71732024-06-27
Gemini 2 Flash Lite Preview63/71792025-02-05
Gemma 2 27B It57/71822024-06-27
Gemini 2.5 Flash Lite187/71942025-06-17
Gemini 2.0 Flash Lite44/72042025-02-25
Gemma 3 27B It106/72132025-03-12
Gemma 2 2B It45/72202024-07-31
Gemma 3 4B34/7243—
Gemma 3 12B45/7248—
Datagemma Rig 27B It13/7—2024-09-12
Diffusiongemma 26B A4B11/7——
Flan T5 Xxl13/7—2022-10-21
Gemini 2 0 Flash Lite Preview 02 0513/7——
Gemini 2 Flash Lite14/7—2025-02-05
Gemini 2.0 Pro11/7——
Gemini 2.0 Pro Exp 02 0521/7——
Gemini 2.5 Flash Lite Preview 09 202511/7——
Gemini 2.5 Flash Preview 09 202511/7——
Gemini 3.1 Pro24/7—2026-02-19
Gemma 1.1 2B It13/7—2024-04-05
Gemma 1.1 7B It24/7—2024-04-05
Gemma 2 9B It Enkrypt Aligned13/7——
Gemma 3 1B11/7——
Gemma 3 270M11/7——
Gemma 3N E2B11/7——
Gemma 3N E4B It22/7—2025-06-03
Gemma 4 12B24/7——
Gemma 4 E2B24/7——
Gemma 4 E4B24/7——
Palm 213/7—2023-05-10
T5 11B11/7——
Ul211/7——
Evaluations covering Google models (70)

AA-Omniscience · AbstentionBench · Adversarial Robustness · Agent-SafetyBench · AgentAbstain · AgentDojo · AILuminate General Purpose AI Chat · AIRBench 2024 Safety Scenarios · Alignment Leaderboard · ANIMA · AnimalHarmBench · Anthropic Agentic Misalignment — blackmail · Anthropic Agentic Misalignment — corporate espionage · Anthropic Agentic Misalignment — lethal action · BioSecBench-Refusal · BrokenMath · BullshitBench v2 · CAIS Risk Index · ChineseSafe · Cisco AI Defense Rolling Single-Turn Leaderboard · Confabulations · DecodingTrust · DSPSafeBench · DystopiaBench · Emergent Collusion · Enkrypt AI Safety Leaderboard · FlagEval Safety and Values · FORTRESS · Google Gemini 2.5 Flash Model Card · Google Gemini 2.5 Flash-Lite Model Card · Gray Swan indirect prompt injection (15 attempts) · HarmBench · HELM Classic RealToxicityPrompts · HELM Safety · HUMAINE Trust, Ethics and Safety · Inkling-Small model card — FORTRESS · Inkling-Small model card — StrongREJECT · Large-scale Moral Machine experiment on LLMs · LiveSecBench · LLM Ethics Benchmark · MACHIAVELLI · Manager Coercion Bench · MANTA · MASK · Microsoft Phi Safety Panels · MORU · ODCV-Bench · OR-Bench · PacifAIst · PandaBench JBB direct-request panel · PHARE · PropensityBench · RefusalBench · S-Eval · SafetyBench · SALAD-Bench · Shell · SM-Bench · Social Welfare Function Benchmark · SORRY-Bench · SOSBench · SpeciesismBench · SpeciEval · SYCON Bench · TAC · ToolPrivacyBench · TrustLLM contemporary collapsed application · UAVBench safety-critical decision recognition · VETO Misfired Alignment · Vigil Mental Health Safety