Claude Haiku 5.5 on BlueBench
Claude Haiku 5.5 ranks 15th of 26 models on BlueBench, Cotool's benchmark for defensive security work, with an overall score of 68.5. It ranks highest in detection engineering (80.5, 9th of 26) and lowest in incident response (65.5, 16th of 26). It averages $0.59 per task, the 7th lowest cost of the 26 models tested.
Anthropic · Closed weight · Updated Oct 2026
Methodology - BlueBench score
- 68.5
- #15 of 26 models
- Cost per task
- $0.59
- #7 cheapest of 26
- Reliability
- 27.9
- #20 of 26 models
- Best-ranked track
- #9
- Detection Engineering · 80.5
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 | 65.5 | #16 of 26 | GPT-6 Astra · 88.0 |
| Threat Hunting | 59.9 | #14 of 26 | Claude Opus 5.5 · 76.1 |
| Detection Engineering | 80.5 | #9 of 26 | GPT-6 Astra · 87.3 |
| Malware Analysis | 67.8 | #20 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 | 64.4 | #15 of 26 |
| Simulation-002Identity & Active Directory Attacks | 72.5 | #10 of 26 |
| Simulation-001Initial Access & Command-and-Control | 54.3 | #11 of 26 |
| Intrusion-003AWS Cloud Intrusion | 73.0 | #16 of 26 |
| Intrusion-002Windows Enterprise Intrusion | 66.7 | #18 of 26 |
| Intrusion-001macOS Threat Investigation | 86.9 | #18 of 26 |
Models ranked nearby
The models closest to Claude Haiku 5.5 on the overall leaderboard.
| Model | Rank | Score | Cost / task |
|---|---|---|---|
| #13 | 70.5 | $0.16 | |
| #14 | 69.9 | $1.01 | |
Claude Haiku 5.5 | #15 | 68.5 | $0.59 |
| #16 | 67.3 | $2.62 | |
| #17 | 65.5 | $1.03 |
Frequently asked questions
- How good is Claude Haiku 5.5 at security operations?
- Claude Haiku 5.5 scores 68.5 overall on BlueBench and ranks 15th 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 Claude Haiku 5.5 best at?
- BlueBench track scores for Claude Haiku 5.5, best rank first. Detection Engineering: 80.5 (9th of 26); Threat Hunting: 59.9 (14th of 26); Incident Response: 65.5 (16th of 26); Malware Analysis: 67.8 (20th of 26).
- How much does Claude Haiku 5.5 cost per security task?
- Claude Haiku 5.5 averages $0.59 per BlueBench task, including repeated attempts. That is the 7th 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.
- How was Claude Haiku 5.5 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