Safety Research

Can your AI be broken?

Name: Rival Jailbreak Safety Benchmark
Creator: Rival

66 models tested across 9 escalating jailbreak levels, from basic prompt injection to multi-step attacks.

Tested66

Resisted6

Avg BreakL4.4

Levels

Safe

Danger

Curious how these models compare beyond safety?

Get the dataset Compare models Full research

It gets worse from here

56+ models cracked. Free. No card. No tricks.

Free account. No card required. By continuing, you agree to Rival's Terms and Privacy Policy

Same 9 tests. 57 models. Wildly different results.

Every prompt, every response, every score.

rival-jailbreak-benchmark.jsonl70 models × 9 levels

claude-opus-..

claude-haiku..

glm-5

nemotron-3-u..

nemotron-3.5..

claude-sonne..

claude-4.5-s..

codex-mini

claude-3.5-s..

minimax-m2.5

gemini-3.1-p..

qwen3.5-35b-..

qwen3.5-27b

qwen3.5-122b..

+54 more models

Get the full dataset$19$9

Download free sample All research

Keep going

CHALLENGES

Watch Models Solve Real Tasks

Code, design, logic. Same task, every model's attempt

COMPARE

Claude 4.5 vs GPT-4.1

The two safest flagships compared on real prompts