EvoRiskBench: An Evolving Benchmark for Runtime Security Risks in Workspace Agents
EvoRiskBench releases 450 adversarial workspace-agent tasks; worst configuration (Codex plus DeepSeek-V4-Pro-0813) reached 68.44% attack success rate.
58
AI model
DeepSeek-V4-Pro-0813
- Mentions · 7d0
- Mentions · 30d1
- All mentions2
- First seenSep 8, 2026
- Last mentioned28d ago
Timeline
2 articles · newest firstnex-agi/Nex-N2.5-Pro — new model trending #30 on Hugging Face
Nex-AGI launches Nex-N2.5 agentic model family (mini/Pro/Max), with Max built on a 1.6-trillion-parameter MoE foundation.
62
Appears with
- Claude Opus 5ai model · 2
- Nex-AGIorganization · 1
- GPT-5.6 Solai model · 1
- Nex-N2.5ai model · 1
- Hugging Faceorganization · 1
- Nex-N2.5-Maxproduct · 1
- OpenRouterorganization · 1
- Nex-AGIvendor · 1
- Kimi-K3ai model · 1
- Nex-N2.5-miniai model · 1
- Codexai model · 1
- GLM-5.3ai model · 1
- ModelScopeorganization · 1
- Nex-N2.5product · 1
- OpenClaworganization · 1
- ModelScopevendor · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.