SecRespond - Baseline Risk Detection: leaderboard
SecRespond (Alibaba and HKUST, 2026) incident-response agents in the OpenCode harness, scored on the checkpoints mapped to the baseline risk capability dimension across 10 compromised cloud-host cyber ranges. The score is the detection score achieved as a percent of the maximum over those checkpoints, averaged over three LLM judges; higher is better.
Metric: Detection CAP-score (%) on the Baseline Risk (BAS) capability dimension: the detection (discovery, evidence and root-cause attribution, 0-3 per checkpoint) score achieved over all checkpoints mapped to the dimension across the 10 SecRespond cyber ranges (averaged over three LLM judges); agents in the OpenCode harness with reasoning enabled; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 (Thinking) | 84.2 | #161 (Claude Sonnet 4.6) |
| 2 | GLM-5.1 | 81.1 | #179 |
| 3 | Qwen 3.7 Max (Thinking) | 80.9 | #145 (Qwen 3.7 Max) |
| 4 | Claude Opus 4.6 (Thinking) | 78.6 | #102 (Claude Opus 4.6) |
| 5 | Claude Opus 4.5 (Thinking) | 75.8 | #156 (Claude Opus 4.5) |
| 6 | Qwen 3.6 Plus (Thinking) | 74.7 | #222 (Qwen 3.6 Plus) |
| 7 | GPT-5.4 Pro (xHigh) | 73.5 | #85 (GPT-5.4 Pro) |
| 8 | MiniMax-M2.7 | 68.6 | #518 |
| 9 | GLM-5 | 67.8 | #260 |
| 10 | Kimi K2.6 (Thinking) | 66.7 | #178 (Kimi K2.6) |
| 11 | Claude Sonnet 4.5 (Thinking) | 66.2 | #250 (Claude Sonnet 4.5) |
| 12 | GPT-5.5 | 63.6 | #49 |
| 13 | GPT-5.2 Pro | 62.4 | #124 |
| 14 | Qwen 3.5 Plus (Thinking) | 61.9 | #232 (Qwen 3.5 Plus) |
| 15 | GPT-5.4 | 61.8 | #123 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=secrespond-baseline-risk-detection · How It Works · Data refreshed daily, snapshot 2026-11-13.