Skip to main content
Benchmark Registry
ModelsBenchmarksOrganizationsCompare

CyberGym

Benchmark
CyberGym
Evaluated by
CyberGym Authors
Release date
June 3, 2025
Version
CyberGym Original
Metric
Vulnerability reproduction success

Results

8 results

LatestHistory
All providersMeta AI (originally Facebook AI Research)SpaceXAIZ.ai
CyberGym Original results
ProviderModelScoreSourceRegistry No.
SpaceXAIGrok 4.7 (high)80.3%Source for Grok 4.7 (high) on CyberGym Original (opens in a new tab)40003
SpaceXAIGrok 4.6 (high)79.7%Source for Grok 4.6 (high) on CyberGym Original (opens in a new tab)40002
Z.aiGLM-5.3 (max)84.5%Source for GLM-5.3 (max) on CyberGym Original (opens in a new tab)150005
Meta AI (originally Facebook AI Research)Muse Spark (high)43.5%Source for Muse Spark (high) on CyberGym Original (opens in a new tab)80001
Meta AI (originally Facebook AI Research)Muse Spark 1.1 (high)59.0%Source for Muse Spark 1.1 (high) on CyberGym Original (opens in a new tab)80002
Z.aiGLM-5.277.2%Source for GLM-5.2 on CyberGym Original (opens in a new tab)150004
Z.aiGLM-5.168.7%Source for GLM-5.1 on CyberGym Original (opens in a new tab)150003
Z.aiGLM-5 (thinking)43.2%Source for GLM-5 (thinking) on CyberGym Original (opens in a new tab)150002
Page 1 of 1

© 2026 Densa Labs

Toggle between the data update date and the application build time.
Legal
Color theme