Skip to main content

Claude Sonnet 5 on BlueBench

Claude Sonnet 5 ranks 26th of 26 models on BlueBench, Cotool's benchmark for defensive security work, with an overall score of 44.8. It ranks highest in detection engineering (67.3, 20th of 26) and lowest in threat hunting (23.8, 26th of 26). It averages $1.77 per task, the 16th lowest cost of the 26 models tested. It refused a material share of tasks on cybersecurity grounds, and its scores include those refusals.

Anthropic · Closed weight · Updated Oct 2026
Methodology
BlueBench score
44.8
#26 of 26 models
Cost per task
$1.77
#16 cheapest of 26
Reliability
8.1
#25 of 26 models
Best-ranked track
#20
Detection Engineering · 67.3
Read the Claude Sonnet 5 rundown →Verdict, strengths and weaknesses against similar-cost models, and when to pick something else.

Scores by track

BlueBench weights incident response, threat hunting, and detection engineering at 30% each and malware analysis at 10%.

TrackScoreRankTrack leader
Incident Response29.2#26 of 26GPT-6 Astra · 88.0
Threat Hunting23.8#26 of 26Claude Opus 5.5 · 76.1
Detection Engineering67.3#20 of 26GPT-6 Astra · 87.3
Malware Analysis86.9#6 of 26Claude Opus 5.5 · 89.4

Scores by benchmark

Each BlueBench benchmark is a separate intrusion, real or simulated, with its own tasks and grading.

Models ranked nearby

The models closest to Claude Sonnet 5 on the overall leaderboard.

ModelRankScoreCost / task
#2256.4$0.90
#2353.6$0.91
#2450.5$4.44
#2545.3$4.63
Claude Sonnet 5
#2644.8$1.77

Frequently asked questions

How good is Claude Sonnet 5 at security operations?
Claude Sonnet 5 scores 44.8 overall on BlueBench and ranks 26th 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 Sonnet 5 best at?
BlueBench track scores for Claude Sonnet 5, best rank first. Malware Analysis: 86.9 (6th of 26); Detection Engineering: 67.3 (20th of 26); Incident Response: 29.2 (26th of 26); Threat Hunting: 23.8 (26th of 26).
How much does Claude Sonnet 5 cost per security task?
Claude Sonnet 5 averages $1.77 per BlueBench task, including repeated attempts. That is the 16th lowest cost of 26 models. The best-scoring cheaper model is Claude Sonnet 5.5 at $1.01 per task with an overall score of 77.5.
Does Claude Sonnet 5 refuse cybersecurity tasks?
Yes. Claude Sonnet 5 refused a material share of BlueBench tasks on cybersecurity grounds. Refused attempts are scored as failures rather than removed, so its scores reflect the refusals.
How was Claude Sonnet 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

Book a demo