CyberGym

CyberGym is a benchmark for evaluating AI agents on cybersecurity tasks, testing their ability to identify vulnerabilities, perform security analysis, and complete security-related challenges in a controlled environment.

17rows
scoreprimary metric
2026-05-28sampled

Metadata

ID
cybergym
Category
Cybersecurity
Release
2025-06-03

Metrics

Score, Normalized Score

Showing 4 latest source slices.

Latest Results

Provider-published system-card benchmark scores parsed from Anthropic's Claude Opus 4.8 capability evaluation tables. Rows are marked self-reported and should be interpreted as source claims unless independently reproduced.

Rank Subject Score Model Match Provenance Sampled
1 Claude Mythos Preview 83.1% Claude Mythos Preview
anthropic-claude-mythos-preview
 		Self-reported 2026-05-28
2 Claude Opus 4.8 78.8% Claude Opus 4.8
anthropic-claude-opus-4.8
 		Self-reported 2026-05-28
3 Claude Opus 4.7 73.1% Claude Opus 4.7
anthropic-claude-opus-4.7
 		Self-reported 2026-05-28
4 Claude Sonnet 4.6 65.2% Claude Sonnet 4.6
anthropic-claude-sonnet-4.6
 		Self-reported 2026-05-28
1 Claude Mythos Preview 0.83 Claude Mythos Preview
anthropic-claude-mythos-preview
 		Self-reported 2026-05-06
2 GPT-5.5 0.82 GPT-5.5
openai-gpt-5.5
 		Self-reported 2026-05-06
3 Claude Opus 4.6 0.74 Claude Opus 4.6
anthropic-claude-opus-4.6
 		Self-reported 2026-05-06
4 Claude Opus 4.7 0.73 Claude Opus 4.7
anthropic-claude-opus-4.7
 		Self-reported 2026-05-06
5 GLM-5.1 0.69 GLM GLM 5.1
z-ai-glm-5.1
 		Self-reported 2026-05-06
6 Kimi K2.5 0.41 KIMI MoonshotAI: Kimi K2.5
moonshotai-kimi-k2.5
 		Self-reported 2026-05-06
1 GPT-5.5 81.8% GPT-5.5
openai-gpt-5.5
 		Launch post 2026-04-23
2 GPT-5.4 79% GPT-5.4
openai-gpt-5.4
 		Launch post 2026-04-23
3 Claude Opus 4.7 73.1% Claude Opus 4.7
anthropic-claude-opus-4.7
 		Launch post 2026-04-23
1 Claude Mythos Preview 83.1% Claude Mythos Preview
anthropic-claude-mythos-preview
 		Launch post 2026-04-16
2 Claude Opus 4.6 73.8% Claude Opus 4.6
anthropic-claude-opus-4.6
 		Launch post 2026-04-16
3 Claude Opus 4.7 73.1% Claude Opus 4.7
anthropic-claude-opus-4.7
 		Launch post 2026-04-16
4 GPT-5.4 66.3% GPT-5.4
openai-gpt-5.4
 		Launch post 2026-04-16