← Back to Research
NexBench
NexBench provides signals of a model's ability to write, reason, and execute real exploits against a real target in a messy environment, while measuring price-to-performance ratio, token efficiency, and more.
Last updated
Highlights
Best Score
Best severity-weighted score · Higher is better
Grok 4.6
Grok 4.7
GPT-5.6 Sol
Grok 4.5
Opus 4.8
GPT-5.6 Luna
Kimi K3
GPT-5.5
GPT-5.6 Terra
GLM-5.2
Sonnet 5
GLM-5.3
GLM-5.3 Flash
Kimi K2.7
Nemotron 3 Ultra
Mercury 2.5 Preview
Best severity-weighted score across every reasoning-effort run= 114
Best severity-weighted score across every reasoning-effort run= 70
Best severity-weighted score across every reasoning-effort run= 61
Best severity-weighted score across every reasoning-effort run= 52
Best severity-weighted score across every reasoning-effort run= 45
Best severity-weighted score across every reasoning-effort run= 43
Best severity-weighted score across every reasoning-effort run= 42
Best severity-weighted score across every reasoning-effort run= 41
Best severity-weighted score across every reasoning-effort run= 36
Best severity-weighted score across every reasoning-effort run= 34
Best severity-weighted score across every reasoning-effort run= 34
Best severity-weighted score across every reasoning-effort run= 33
Best severity-weighted score across every reasoning-effort run= 29
Best severity-weighted score across every reasoning-effort run= 24
Best severity-weighted score across every reasoning-effort run= 18
Best severity-weighted score across every reasoning-effort run= 6
075150
114
70
61
52
45
43
42
41
36
34
34
33
29
24
18
6
Best Runtime
Best-run wall time · Lower is better
Kimi K3
Mercury 2.5 Preview
GPT-5.6 Terra
GLM-5.3
Grok 4.7
GPT-5.5
GLM-5.2
Opus 4.8
Grok 4.6
GPT-5.6 Luna
GPT-5.6 Sol
Kimi K2.7
GLM-5.3 Flash
Sonnet 5
Grok 4.5
Nemotron 3 Ultra
Wall-clock run time of the best-scoring attempt= 49.0m
Wall-clock run time of the best-scoring attempt= 66.0m
Wall-clock run time of the best-scoring attempt= 66.2m
Wall-clock run time of the best-scoring attempt= 77.0m
Wall-clock run time of the best-scoring attempt= 77.0m
Wall-clock run time of the best-scoring attempt= 78.8m
Wall-clock run time of the best-scoring attempt= 100.0m
Wall-clock run time of the best-scoring attempt= 110.5m
Wall-clock run time of the best-scoring attempt= 117.0m
Wall-clock run time of the best-scoring attempt= 119.0m
Wall-clock run time of the best-scoring attempt= 139.1m
Wall-clock run time of the best-scoring attempt= 152.0m
Wall-clock run time of the best-scoring attempt= 204.0m
Wall-clock run time of the best-scoring attempt= 269.7m
Wall-clock run time of the best-scoring attempt= 302.5m
Wall-clock run time of the best-scoring attempt= 302.7m
0.0m175.0m350.0m
49.0m
66.0m
66.2m
77.0m
77.0m
78.8m
100.0m
110.5m
117.0m
119.0m
139.1m
152.0m
204.0m
269.7m
302.5m
302.7m
Cost per Finding
Standard cost per accepted finding · Lower is better
GLM-5.3 Flash
Grok 4.6
Kimi K2.7
Kimi K3
GLM-5.3
GLM-5.2
Grok 4.7
Grok 4.5
Mercury 2.5 Preview
GPT-5.5
GPT-5.6 Luna
GPT-5.6 Terra
Opus 4.8
Nemotron 3 Ultra
Sonnet 5
GPT-5.6 Sol
$19 standard cost ÷ 22 accepted findings= $0.85
$39 standard cost ÷ 44 accepted findings= $0.88
$14 standard cost ÷ 14 accepted findings= $0.99
$47 standard cost ÷ 41 accepted findings= $1.15
$30 standard cost ÷ 25 accepted findings= $1.21
$38 standard cost ÷ 31 accepted findings= $1.22
$44 standard cost ÷ 26 accepted findings= $1.67
$122 standard cost ÷ 51 accepted findings= $2.39
$7 standard cost ÷ 3 accepted findings= $2.48
$120 standard cost ÷ 45 accepted findings= $2.67
$158 standard cost ÷ 47 accepted findings= $3.36
$215 standard cost ÷ 49 accepted findings= $4.38
$380 standard cost ÷ 68 accepted findings= $5.59
$92 standard cost ÷ 16 accepted findings= $5.77
$374 standard cost ÷ 54 accepted findings= $6.93
$1093 standard cost ÷ 87 accepted findings= $12.56
$0.00$7.50$15.00
$0.85
$0.88
$0.99
$1.15
$1.21
$1.22
$1.67
$2.39
$2.48
$2.67
$3.36
$4.38
$5.59
$5.77
$6.93
$12.56
Cost Efficiency
Severity-weighted points per dollar · Higher is better
Grok 4.6
GLM-5.3 Flash
GLM-5.3
Kimi K3
Kimi K2.7
GLM-5.2
Grok 4.7
Grok 4.5
GPT-5.5
Mercury 2.5 Preview
GPT-5.6 Luna
GPT-5.6 Terra
Opus 4.8
Sonnet 5
Nemotron 3 Ultra
GPT-5.6 Sol
114 points ÷ $39 standard cost= 2.95
50 points ÷ $19 standard cost= 2.66
62 points ÷ $30 standard cost= 2.04
94 points ÷ $47 standard cost= 2.00
24 points ÷ $14 standard cost= 1.72
65 points ÷ $38 standard cost= 1.72
70 points ÷ $44 standard cost= 1.61
118 points ÷ $122 standard cost= 0.97
98 points ÷ $120 standard cost= 0.82
6 points ÷ $7 standard cost= 0.81
100 points ÷ $158 standard cost= 0.63
105 points ÷ $215 standard cost= 0.49
157 points ÷ $380 standard cost= 0.41
130 points ÷ $374 standard cost= 0.35
27 points ÷ $92 standard cost= 0.29
192 points ÷ $1093 standard cost= 0.18
0.001.503.00
2.95
2.66
2.04
2.00
1.72
1.72
1.61
0.97
0.82
0.81
0.63
0.49
0.41
0.35
0.29
0.18
Token Efficiency
Severity-weighted points per 1M tokens · Higher is better
Grok 4.6
Grok 4.5
Kimi K3
Opus 4.8
Grok 4.7
GLM-5.3
GLM-5.2
Kimi K2.7
GPT-5.6 Terra
GLM-5.3 Flash
GPT-5.5
GPT-5.6 Sol
GPT-5.6 Luna
Sonnet 5
Mercury 2.5 Preview
Nemotron 3 Ultra
114 points ÷ 25.4M tokens= 4.48
118 points ÷ 49.6M tokens= 2.38
94 points ÷ 40.0M tokens= 2.35
157 points ÷ 77.8M tokens= 2.02
70 points ÷ 35.3M tokens= 1.99
62 points ÷ 40.6M tokens= 1.53
65 points ÷ 42.7M tokens= 1.52
24 points ÷ 17.3M tokens= 1.39
105 points ÷ 80.3M tokens= 1.31
50 points ÷ 39.4M tokens= 1.27
98 points ÷ 88.0M tokens= 1.11
192 points ÷ 224.5M tokens= 0.86
100 points ÷ 127.8M tokens= 0.78
130 points ÷ 173.1M tokens= 0.75
6 points ÷ 20.1M tokens= 0.30
27 points ÷ 261.5M tokens= 0.10
0.002.254.50
4.48
2.38
2.35
2.02
1.99
1.53
1.52
1.39
1.31
1.27
1.11
0.86
0.78
0.75
0.30
0.10
Full standings
Model leaderboard
Best score and run time come from one run. Findings, cost, and tokens are totals across the number of runs shown. Ranked by best score; click any column header to re-sort.
| Model | Best-run effort | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Grok 4.6 | X-High | 114 | 1 | 1h 57m | 44 | 26 | $39 | 25.4M | 2.951 | 4.485 |
| Grok 4.7 | X-High | 70 | 1 | 1h 17m | 26 | 20 | $44 | 35.3M | 1.609 | 1.986 |
| GPT-5.6 Sol | X-High | 61 | 5 | 2h 19m | 87 | 40 | $1,093 | 224.5M | 0.176 | 0.855 |
| Grok 4.5 | High | 52 | 3 | 5h 3m | 51 | 26 | $122 | 49.6M | 0.970 | 2.378 |
| Claude Opus 4.8 | Max | 45 | 5 | 1h 50m | 68 | 38 | $380 | 77.8M | 0.413 | 2.019 |
| GPT-5.6 Luna | X-High | 43 | 5 | 1h 59m | 47 | 21 | $158 | 127.8M | 0.633 | 0.782 |
| Kimi K3 | Default | 42 | 3 | 49m | 41 | 22 | $47 | 40.0M | 2.000 | 2.349 |
| GPT-5.5 | X-High | 41 | 5 | 1h 19m | 45 | 19 | $120 | 88.0M | 0.817 | 1.114 |
| GPT-5.6 Terra | X-High | 36 | 5 | 1h 6m | 49 | 19 | $215 | 80.3M | 0.489 | 1.308 |
| GLM-5.2 | X-High | 34 | 2 | 1h 40m | 31 | 12 | $38 | 42.7M | 1.722 | 1.522 |
| Claude Sonnet 5 | Max | 34 | 5 | 4h 30m | 54 | 31 | $374 | 173.1M | 0.348 | 0.751 |
| GLM-5.3 | Max | 33 | 2 | 1h 17m | 25 | 15 | $30 | 40.6M | 2.044 | 1.528 |
| GLM-5.3 Flash | Max | 29 | 2 | 3h 24m | 22 | 11 | $19 | 39.4M | 2.661 | 1.268 |
| Kimi K2.7 | Default | 24 | 1 | 2h 32m | 14 | 2 | $14 | 17.3M | 1.724 | 1.387 |
| Nemotron 3 Ultra | High | 18 | 2 | 5h 3m | 16 | 2 | $92 | 261.5M | 0.292 | 0.103 |
| Mercury 2.5 Preview | High | 6 | 1 | 1h 6m | 3 | 1 | $7 | 20.1M | 0.805 | 0.299 |
Accepted findings by severity
Validator-accepted findings per model, sorted by total volume. Each model keeps its own hue; darker shades are higher-severity findings. Hover a bar for the low / medium / high split.
Hue = model · shade = severity
High / critical
Medium
Low
GPT-5.6 SolX-HighOpenAI
Opus 4.8MaxAnthropic
Sonnet 5MaxAnthropic
Grok 4.5HighxAI
GPT-5.6 TerraX-HighOpenAI
GPT-5.6 LunaX-HighOpenAI
GPT-5.5X-HighOpenAI
Grok 4.6X-HighxAI
Kimi K3DefaultMoonshot AI
GLM-5.2X-HighZ.ai
Grok 4.7X-HighxAI
GLM-5.3MaxZ.ai
GLM-5.3 FlashMaxZ.ai
Nemotron 3 UltraHighNVIDIA
Kimi K2.7DefaultMoonshot AI
Mercury 2.5 PreviewHighInception
87
high40
medium25
low22
87 accepted findings
68
high38
medium13
low17
68 accepted findings
54
high31
medium14
low9
54 accepted findings
51
high26
medium15
low10
51 accepted findings
49
high19
medium18
low12
49 accepted findings
47
high21
medium11
low15
47 accepted findings
45
high19
medium15
low11
45 accepted findings
44
high26
medium18
low0
44 accepted findings
41
high22
medium9
low10
41 accepted findings
31
high12
medium10
low9
31 accepted findings
26
high20
medium4
low2
26 accepted findings
25
high15
medium7
low3
25 accepted findings
22
high11
medium6
low5
22 accepted findings
16
high2
medium7
low7
16 accepted findings
14
high2
medium6
low6
14 accepted findings
3
high1
medium1
low1
3 accepted findings
0255075100