NUWA 
DeepSeek
DeepSeek R1-0528
Published evidence on capabilities, behavioral tendencies and observed actions.
- First public availability
- 2025-05-28
- Parameters
- Not yet verified
- Metadata updated
- 2026-10-27
Evidence overview
Published risk evidence is organized by assessment and protocol. Missing evidence is not a low-risk finding.
NUWA curates and interprets these sources. Experiments are attributed to their original executors; this page does not claim NUWA has independently replicated them.
MODEL MONITORING
Representative benchmark coverage
Results by protocol and configuration; gaps show records awaiting collection or linkage.
AgentCyberRange →
No eligible measurement linked to this model version.
Cybench →
1 measurements
- 35.5%NIST / CAISI · CAISI ReAct-style agent; Inspect
CyberGym →
No eligible measurement linked to this model version.
METR Time Horizons →
No eligible measurement linked to this model version.
SEC-bench Pro →
No eligible measurement linked to this model version.
MASK →
No eligible measurement linked to this model version.
Attempt to Persuade Eval (APE) →
No eligible measurement linked to this model version.
DeceptionBench →
No eligible measurement linked to this model version.
PersuSafety →
No eligible measurement linked to this model version.
All linked measurements
Each result retains its protocol and unit.
4 observations
| System / model | Result | Condition status | Task / conditions | Evaluator / report | Dates |
|---|---|---|---|---|---|
| DeepSeek R1-0528 | 37%AgentDojo · Hijacking attempted-task frequency | Incomplete | AgentDojo Workspace; Credential exfiltrationHijacking attempted-task frequency— | NIST / CAISIEvaluation of DeepSeek AI Models | Release 2025-05-28Report 2025-09-30 |
| DeepSeek R1-0528 | 48%AgentDojo · Hijacking attempted-task frequency | Incomplete | AgentDojo Workspace; Phishing emailHijacking attempted-task frequency— | NIST / CAISIEvaluation of DeepSeek AI Models | Release 2025-05-28Report 2025-09-30 |
| DeepSeek R1-0528 | 49%AgentDojo · Hijacking attempted-task frequency | Incomplete | AgentDojo Workspace; Download/run malwareHijacking attempted-task frequency— | NIST / CAISIEvaluation of DeepSeek AI Models | Release 2025-05-28Report 2025-09-30 |
| DeepSeek R1-0528 | 35.5%Cybench · Reported benchmark performance | Incomplete | CybenchReported benchmark performance500,000 weighted tokens per agentic task (Appendix A2)CAISI ReAct-style agent; Inspect | NIST / CAISIEvaluation of DeepSeek AI Models | Release 2025-05-28Report 2025-09-30 |
Developer disclosures
No developer-authored safety report has been linked to this exact model record in the current collection.
External assessments
NIST / CAISI
Evaluation of DeepSeek AI Models →
Reported differences depend on task, scaffold and scoring; institutional judgments are retained as source judgments rather than NUWA findings.
2025-09-30Evidence by risk topic
Autonomy No structured result linked
A gap in this collection is not evidence of low risk. Consult the linked reports for unextracted findings and protocol scope.
Deception No structured result linked
A gap in this collection is not evidence of low risk. Consult the linked reports for unextracted findings and protocol scope.
Cybersecurity 4 protocol results
Cybench →
External evaluation · Executed by NIST / CAISI
- Reported benchmark performance
- 35.5%
Source denominator: 40 (unit not disclosed)
Agent hijacking →
External evaluation · Executed by NIST / CAISI
- Hijacking attempted-task frequency
- 37%
Per-result sample size not disclosed
Agent hijacking →
External evaluation · Executed by NIST / CAISI
- Hijacking attempted-task frequency
- 48%
Per-result sample size not disclosed
Agent hijacking →
External evaluation · Executed by NIST / CAISI
- Hijacking attempted-task frequency
- 49%
Per-result sample size not disclosed
CBRN No structured result linked
A gap in this collection is not evidence of low risk. Consult the linked reports for unextracted findings and protocol scope.
Disinformation No structured result linked
A gap in this collection is not evidence of low risk. Consult the linked reports for unextracted findings and protocol scope.
Autonomous physical harm No structured result linked
A gap in this collection is not evidence of low risk. Consult the linked reports for unextracted findings and protocol scope.
Source review and evidence gaps
Evaluation of DeepSeek AI Models →
Source-specific extraction of CAISI · DeepSeek models. Findings and measurements retain their original attribution.
Source transcription is not experimental replication. Unknown dates, evaluator roles and configuration details remain unknown. Source groups are not verified rankings.
Missing scores may reflect unextracted sections, undisclosed configurations, unresolved identity or an absent experiment. These are separate research questions; none implies a zero result.
Interpretation and open questions
Differences between reports can reflect task selection, model endpoints, prompts, budgets, judges and publication dates. Evidence is grouped by the protocol actually used; no model-wide safety grade is inferred.
References
- [1] deepseek-r1-0528 release ; News article title, dated heading and availability
- [2] Evaluation of DeepSeek AI Models. 2025-09-30. 30 September 2025 ; p. 24, Table 4.1, row Cybench, column DeepSeek R1-0528; p. 17, key findings, Credential exfiltration; p. 17, key findings, Phishing email