Skip to main content

DeepSeek V4 Flash security benchmark rundown

DeepSeek V4 Flash (DeepSeek) ranks 21st of 26 on BlueBench with 58.6 overall at $0.17 per task. Track scores, value against peers, strengths, weaknesses, and when to use it.

DeepSeek · Open weight · Tested Sep 11, 2026 · Data updated Oct 2026
Methodology

Verdict

DeepSeek V4 Flash ranks 21st of 26 on BlueBench with an overall score of 58.6, in the bottom quarter of the leaderboard and 24.6 points behind Claude Opus 5.5. At $0.17 per task it is the 3rd cheapest of 26, and GLM-5.3 Flash scores higher (70.6) for less ($0.11). Against the 4 models priced within 2x of its cost per task, it scores 11.9 points below the peer median overall; it trails them most in incident response. Malware Analysis, weighted 10%, is a secondary weak spot (−11.6 vs the peer median). Its results vary more than most across the benchmarks (reliability 23.7, 22nd of 26).

BlueBench score
58.6
#21 of 26 models
Cost per task
$0.17
#3 cheapest of 26
Reliability
23.7
#22 of 26 models
Overall vs cost peers
−11.9
Off the cost–score frontier
All DeepSeek V4 Flash results and FAQ →

Capability by track

Incident response, threat hunting, and detection engineering each carry 30% of the BlueBench score; malware analysis carries 10%.

DeepSeek V4 Flash scores 58.6 overall, 21st of 26 on the capability leaderboard. Its best rank is in incident response (57.7, 20th of 26) and its worst is in malware analysis (62.8, 22nd of 26). Among the 8 open-weight models tested it ranks 6th.

TrackWeightScoreRankPeer medianvs peers
Incident Response30%57.7#20 of 2671.8−14.1
Threat Hunting30%51.9#21 of 2661.4−9.5
Detection Engineering30%64.8#22 of 2673.6−8.8
Malware Analysis10%62.8#22 of 2674.4−11.6

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.

DeepSeek V4 Flash averages $0.17 per task, including repeated attempts: the 3rd cheapest of 26 models with a cost figure. It is off the cost–score frontier: GLM-5.3 Flash scores 70.6 at $0.11, and 1 other model is also both cheaper and at least as good. Its peer group is the 4 models priced within 2x of its cost per task. Their median overall score is 70.5, so DeepSeek V4 Flash is 11.9 points behind its price range. Reliability, 100 minus its mean score range across the benchmarks, is 23.7 (22nd of 26).

Strengths and weaknesses

Each track compared with the median of the 4 models priced within 2x of its cost per task. A gap of 5 points or more counts.

Against the 4 models priced within 2x of its cost per task, DeepSeek V4 Flash lags in incident response (−14.1), threat hunting (−9.5), detection engineering (−8.8) and malware analysis (−11.6). Differences are versus the peer median score in each track; ±5 points is the threshold.

Beats its cost peers

No track by 5+ points

Lags its cost peers

  • Incident Response57.7 (−14.1)
  • Threat Hunting51.9 (−9.5)
  • Detection Engineering64.8 (−8.8)
  • Malware Analysis62.8 (−11.6)

Nearest neighbors on the leaderboard

The models ranked just above and below it overall, with the largest track gaps.

  • Gemini 3.6 Flash sits 2 places above, 4.1 points higher overall (62.7), at 17.6x the cost per task ($3.01); it is 16.3 points ahead in detection engineering.
  • GPT-5.6 Luna sits one place above, 3.8 points higher overall (62.4), at 3.4x the cost per task ($0.58); it is 16.3 points ahead in malware analysis.
  • MiniMax M3 sits one place below, 2.2 points lower overall (56.4), at 5.2x the cost per task ($0.90); it is 6.3 points ahead in malware analysis and 8.0 points behind in detection engineering.
  • DeepSeek V4 Pro sits 2 places below, 5.0 points lower overall (53.6), at 5.3x the cost per task ($0.91); it is 10.6 points behind in threat hunting.
RankModelScore$/task
19Gemini 3.6 Flash62.7$3.01
20GPT-5.6 Luna62.4$0.58
21DeepSeek V4 Flash58.6$0.17
22MiniMax M356.4$0.90
23DeepSeek V4 Pro53.6$0.91

Best for

Alternatives are limited to models costing at most 1.5x as much, except the overall leader and the cheaper-model pick.

  • Use itBlueBench does not single out a task for it: it is in the top 3 of no track and beats its cost peers by 5+ points in none.
  • InsteadUse GLM-5.3 Flash instead in most cases: it scores 70.6 overall (+12.0) at $0.11 per task (62% of the cost per task).
  • InsteadUse GPT-6.1 Terra instead if incident response is the priority: it scores 78.2 there (vs 57.7) at $0.16 per task.
  • InsteadUse Claude Opus 5.5 instead if the highest overall score matters more than cost: it scores 83.2 (24.6 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.

  • Its best standing is on Simulation-003 (Impact & Exfiltration): 63.1 (17th of 26).
  • Its scores range from 44.3 on Simulation-002 to 86.8 on Intrusion-001, a spread of 42.5 points.
  • 1 recorded refusal across the benchmarks; refused attempts stay in the score.
BenchmarkScoreRank
Simulation-003Impact & Exfiltration63.1#17 of 26
Simulation-002Identity & Active Directory Attacks44.3#23 of 26
Simulation-001Initial Access & Command-and-Control45.9#20 of 26
Intrusion-003AWS Cloud Intrusion56.5#22 of 26
Intrusion-002Windows Enterprise Intrusion62.9#19 of 26
Intrusion-001macOS Threat Investigation86.8#19 of 26

About this rundown

Provider
DeepSeek · open weight
Tested
Sep 11, 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

Book a demo