Grok 4.6 on BlueBench
Grok 4.6 ranks 4th of 26 models on BlueBench, Cotool's benchmark for defensive security work, with an overall score of 79.2. It ranks highest in threat hunting (72.6, 3rd of 26) and lowest in incident response (80.8, 7th of 26). It averages $3.64 per task, the 22nd lowest cost of the 26 models tested.
xAI · Closed weight · Updated Oct 2026
Methodology - BlueBench score
- 79.2
- #4 of 26 models
- Cost per task
- $3.64
- #22 cheapest of 26
- Reliability
- 47.4
- #6 of 26 models
- Best-ranked track
- #3
- Threat Hunting · 72.6
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 | 80.8 | #7 of 26 | GPT-6 Astra · 88.0 |
| Threat Hunting | 72.6 | #3 of 26 | Claude Opus 5.5 · 76.1 |
| Detection Engineering | 82.6 | #5 of 26 | GPT-6 Astra · 87.3 |
| Malware Analysis | 83.8 | #7 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 | 80.6 | #6 of 26 |
| Simulation-002Identity & Active Directory Attacks | 80.8 | #5 of 26 |
| Simulation-001Initial Access & Command-and-Control | 64.9 | #4 of 26 |
| Intrusion-003AWS Cloud Intrusion | 87.4 | #4 of 26 |
| Intrusion-002Windows Enterprise Intrusion | 72.1 | #10 of 26 |
| Intrusion-001macOS Threat Investigation | 97.5 | #4 of 26 |
Models ranked nearby
The models closest to Grok 4.6 on the overall leaderboard.
| Model | Rank | Score | Cost / task |
|---|---|---|---|
| #2 | 80.3 | $9.60 | |
| #3 | 79.9 | $6.76 | |
Grok 4.6 | #4 | 79.2 | $3.64 |
| #5 | 77.5 | $1.01 | |
| #6 | 77.2 | $2.89 |
Frequently asked questions
- How good is Grok 4.6 at security operations?
- Grok 4.6 scores 79.2 overall on BlueBench and ranks 4th 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 Grok 4.6 best at?
- BlueBench track scores for Grok 4.6, best rank first. Threat Hunting: 72.6 (3rd of 26); Detection Engineering: 82.6 (5th of 26); Malware Analysis: 83.8 (7th of 26); Incident Response: 80.8 (7th of 26).
- How much does Grok 4.6 cost per security task?
- Grok 4.6 averages $3.64 per BlueBench task, including repeated attempts. That is the 22nd lowest cost of 26 models. The best-scoring cheaper model is Claude Opus 5.5 at $3.54 per task with an overall score of 83.2.
- How was Grok 4.6 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