Claude Sonnet 5.5 security benchmark rundown
Claude Sonnet 5.5 (Anthropic) ranks 5th of 26 on BlueBench with 77.5 overall at $1.01 per task. Track scores, value against peers, strengths, weaknesses, and when to use it.
Verdict
Claude Sonnet 5.5 ranks 5th of 26 on BlueBench with an overall score of 77.5, in the top quarter of the leaderboard and 5.7 points behind Claude Opus 5.5. At $1.01 per task it is the 11th cheapest of 26, and it sits on the cost–score frontier: no tested model is both cheaper and higher scoring. Against the 10 models priced within 2x of its cost per task, it scores 10.5 points above the peer median overall; it is strongest against them in incident response. Malware Analysis, weighted 10%, is a secondary strength (+11.4 vs the peer median). Its results are among the most consistent across the benchmarks (reliability 50.0, 4th of 26).
- BlueBench score
- 77.5
- #5 of 26 models
- Cost per task
- $1.01
- #11 cheapest of 26
- Reliability
- 50.0
- #4 of 26 models
- Overall vs cost peers
- +10.5
- On the cost–score frontier
Capability by track
Incident response, threat hunting, and detection engineering each carry 30% of the BlueBench score; malware analysis carries 10%.
Claude Sonnet 5.5 scores 77.5 overall, 5th of 26 on the capability leaderboard. Its best rank is in malware analysis (87.8, 3rd of 26) and its worst is in detection engineering (78.1, 12th of 26).
| Track | Weight | Score | Rank | Peer median | vs peers |
|---|---|---|---|---|---|
| Incident Response | 30% | 83.9 | #5 of 26 | 64.2 | +19.8 |
| Threat Hunting | 30% | 66.9 | #7 of 26 | 55.0 | +11.9 |
| Detection Engineering | 30% | 78.1 | #12 of 26 | 76.7 | +1.4 |
| Malware Analysis | 10% | 87.8 | #3 of 26 | 76.4 | +11.4 |
Cost, reliability, and value
Cost is USD per evaluated task including repeated attempts. Reliability is 100 minus the mean score range across the six benchmarks.
Claude Sonnet 5.5 averages $1.01 per task, including repeated attempts: the 11th cheapest of 26 models with a cost figure. No tested model is both cheaper and higher scoring, so it is on the cost–score frontier. Its peer group is the 10 models priced within 2x of its cost per task. Their median overall score is 67.0, so Claude Sonnet 5.5 is 10.5 points ahead of its price range. Reliability, 100 minus its mean score range across the benchmarks, is 50.0 (4th of 26).
Cost peers: the 10 models priced within 2x of its cost per task
Strengths and weaknesses
Each track compared with the median of the 10 models priced within 2x of its cost per task. A gap of 5 points or more counts.
Against the 10 models priced within 2x of its cost per task, Claude Sonnet 5.5 is strongest in incident response (+19.8), threat hunting (+11.9) and malware analysis (+11.4). Differences are versus the peer median score in each track; ±5 points is the threshold.
Beats its cost peers
- Incident Response83.9 (+19.8)
- Threat Hunting66.9 (+11.9)
- Malware Analysis87.8 (+11.4)
Lags its cost peers
No track by 5+ points
Nearest neighbors on the leaderboard
The models ranked just above and below it overall, with the largest track gaps.
- Gemini 4 Argon sits 2 places above, 2.4 points higher overall (79.9), at 6.7x the cost per task ($6.76); it is 7.6 points ahead in threat hunting and 15.3 points behind in malware analysis.
- Grok 4.6 sits one place above, 1.7 points higher overall (79.2), at 3.6x the cost per task ($3.64); it is 5.7 points ahead in threat hunting.
- GPT-5.6 Sol sits one place below, 0.3 points lower overall (77.2), at 2.9x the cost per task ($2.89); no track differs by 5 points or more.
- GPT-6.1 Sol sits 2 places below, 0.8 points lower overall (76.7), at 1.3x the cost per task ($1.30); it is 5.2 points behind in threat hunting.
| Rank | Model | Score | $/task |
|---|---|---|---|
| 3 | Gemini 4 Argon | 79.9 | $6.76 |
| 4 | Grok 4.6 | 79.2 | $3.64 |
| 5 | Claude Sonnet 5.5 | 77.5 | $1.01 |
| 6 | GPT-5.6 Sol | 77.2 | $2.89 |
| 7 | GPT-6.1 Sol | 76.7 | $1.30 |
Best for
Alternatives are limited to models costing at most 1.5x as much, except the overall leader and the cheaper-model pick.
- Use itUse it for malware analysis (87.8, 3rd of 26): it is a top-3 result among all models.
- Use itUse it for incident response (83.9, +19.8) and threat hunting (66.9, +11.9) at its price: those tracks beat the median of the 10 models priced within 2x of its cost per task by 5 points or more.
- Use itUse it when the budget is about $1.01 per task: no model at or below that cost scores higher overall.
- InsteadUse DeepSeek V4.1 Flash instead if cost per task matters most: it scores 71.5 (6.0 points lower) at $0.17, 17% of the cost per task.
- InsteadUse Claude Opus 5.5 instead if the highest overall score matters more than cost: it scores 83.2 (5.7 points higher) at $3.54 per task.
Notable benchmark results
Each BlueBench benchmark is a separate intrusion, real or simulated, with its own tasks and grading.
- 3rd place on Simulation-003 (Impact & Exfiltration): 83.4 (3rd of 26).
- Its scores range from 63.6 on Simulation-001 to 92.8 on Intrusion-001, a spread of 29.2 points.
| Benchmark | Score | Rank |
|---|---|---|
| Simulation-003Impact & Exfiltration | 83.4 | #3 of 26 |
| Simulation-002Identity & Active Directory Attacks | 83.1 | #4 of 26 |
| Simulation-001Initial Access & Command-and-Control | 63.6 | #5 of 26 |
| Intrusion-003AWS Cloud Intrusion | 67.1 | #19 of 26 |
| Intrusion-002Windows Enterprise Intrusion | 78.3 | #4 of 26 |
| Intrusion-001macOS Threat Investigation | 92.8 | #7 of 26 |
Notes
Written by hand; they only restate the data above.
- Of the five highest-ranked models, Claude Sonnet 5.5 has the lowest cost per task ($1.01).
About this rundown
- Provider
- Anthropic · closed weight
- Tested
- Sep 30, 2026 – Oct 1, 2026
- Suite
- 74 tasks across 6 benchmarks, evaluated in a common harness
- Data updated
- Oct 27, 2026
- How it is written
- Generated from the BlueBench result files with fixed rules: cost peers are models priced within 2x per task (or the nearest by cost when fewer than three are), a track is a strength or weakness at 5+ points from the peer median, and ranks share a place on ties. The text is regenerated whenever the results change.
AI for the blue team.
Run Cotool's harness in your environment to get real security work done