DeepSeek V4 Pro on BlueBench
DeepSeek V4 Pro ranks 23rd of 26 models on BlueBench, Cotool's benchmark for defensive security work, with an overall score of 53.6. It ranks highest in detection engineering (65.6, 21st of 26) and lowest in threat hunting (41.3, 24th of 26). It averages $0.91 per task, the 9th lowest cost of the 26 models tested. It is the 8th highest-scoring of 8 open-weight models.
DeepSeek · Open weight · Updated Oct 2026
Methodology - BlueBench score
- 53.6
- #23 of 26 models
- Cost per task
- $0.91
- #9 cheapest of 26
- Reliability
- 31.1
- #18 of 26 models
- Best-ranked track
- #21
- Detection Engineering · 65.6
Scores by track
BlueBench weights incident response, threat hunting, and detection engineering at 30% each and malware analysis at 10%.
| Track | Score | Rank | Track leader |
|---|---|---|---|
| Incident Response | 51.6 | #23 of 26 | GPT-6 Astra · 88.0 |
| Threat Hunting | 41.3 | #24 of 26 | Claude Opus 5.5 · 76.1 |
| Detection Engineering | 65.6 | #21 of 26 | GPT-6 Astra · 87.3 |
| Malware Analysis | 60.0 | #24 of 26 | Claude Opus 5.5 · 89.4 |
Scores by benchmark
Each BlueBench benchmark is a separate intrusion, real or simulated, with its own tasks and grading.
| Benchmark | Score | Rank |
|---|---|---|
| Simulation-003Impact & Exfiltration | 56.5 | #22 of 26 |
| Simulation-002Identity & Active Directory Attacks | 40.9 | #24 of 26 |
| Simulation-001Initial Access & Command-and-Control | 23.6 | #26 of 26 |
| Intrusion-003AWS Cloud Intrusion | 60.0 | #21 of 26 |
| Intrusion-002Windows Enterprise Intrusion | 60.3 | #21 of 26 |
| Intrusion-001macOS Threat Investigation | 90.2 | #12 of 26 |
Models ranked nearby
The models closest to DeepSeek V4 Pro on the overall leaderboard.
| Model | Rank | Score | Cost / task |
|---|---|---|---|
| #21 | 58.6 | $0.17 | |
| #22 | 56.4 | $0.90 | |
DeepSeek V4 Pro | #23 | 53.6 | $0.91 |
| #24 | 50.5 | $4.44 | |
| #25 | 45.3 | $4.63 |
More DeepSeek models
Frequently asked questions
- How good is DeepSeek V4 Pro at security operations?
- DeepSeek V4 Pro scores 53.6 overall on BlueBench and ranks 23rd of 26 models on incident response, threat hunting, detection engineering, and malware analysis. The top model, Claude Opus 5.5, scores 83.2.
- Which security tasks is DeepSeek V4 Pro best at?
- BlueBench track scores for DeepSeek V4 Pro, best rank first. Detection Engineering: 65.6 (21st of 26); Incident Response: 51.6 (23rd of 26); Malware Analysis: 60.0 (24th of 26); Threat Hunting: 41.3 (24th of 26).
- How much does DeepSeek V4 Pro cost per security task?
- DeepSeek V4 Pro averages $0.91 per BlueBench task, including repeated attempts. That is the 9th lowest cost of 26 models. The best-scoring cheaper model is DeepSeek V4.1 Flash at $0.17 per task with an overall score of 71.5.
- Is DeepSeek V4 Pro the best open-weight model for security work?
- No. DeepSeek V4 Pro ranks 8th of 8 open-weight models. DeepSeek V4.1 Flash leads the open-weight models with 71.5.
- How was DeepSeek V4 Pro tested?
- Every model runs the same 74 tasks across 6 BlueBench benchmarks, built from real intrusion logs and simulated enterprise environments, using the same harness, prompts, tools, and number of trials. Reports are graded against hidden rubrics and detection rules are re-run against the evidence.
Full leaderboards
AI for the blue team.
Run Cotool's harness in your environment to get real security work done