Model results
CyberGym
Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.
Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.
Reviewed recommendation and limits
Start with CyberGym Level 1 for reproducing a described vulnerability in an unpatched codebase. Compare matching agent settings and trial budgets.
Level 1. Score the agent-selected final PoC against vulnerable and patched builds. Keep success across multiple trials separate from final-submission scoring within a trial. Preserve dynamic access, cross-instance memory, network policy and agent budget.. Imported groups below may use different protocols; this review does not certify each run. Reproduction does not establish novel vulnerability discovery or exploit development. Network access and exposed patches can leak answers. Imported vendor-reported rows with unknown harnesses are not certified as the reviewed protocol.
CyberGym
- Source
- DeepSeek model card (vendor-reported)
- Version
- unversioned
- Measurement date
- 2026-09-10
- Metric
- % · Higher is better
- Evaluation scope
- Vendor-reported; reasoning effort 100; Pass@1; temperature 1.0, top_p 0.95; Harness not specified in the comparison table
Positions are withheld because at least one result lacks explicit independent provenance or a documented evaluation scope. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.
DeepSeek V4.1 Flash
deepseek/deepseek-v4.1-flash-20260910
88.1%
View modelEvidence model-card-reported