safety benchmark
CVE-Bench Leaderboard
CVE-Bench evaluates agents on exploiting real-world web-application CVEs in sandboxed environments that mimic production services.
Updated Sep 24, 2026
Models1
Model coverage1
MetricScore
EvidenceB
CVE-Bench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelXA | Score36.60% | Percentile100.00% | Participants1 | EvidenceC | Evaluated |
CVE-Bench Highlights
The leading models and scores on this benchmark.
The Top AI Models for CVE-Bench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis cve-bench AI model leaderboard uses descending score in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01XAxAI
- Score
- 36.60%
- Price
- $2.0 input / $6.0 output per 1M tokens
- Speed
- Up to 5.32 tok/s via xAI
Strengths
- Ranks #1 of 1 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures CVE-Bench, not total model capability
Selection summary
Best AI Models for CVE-Bench
Grok 4.7 currently leads CVE-Bench with 36.60%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price, and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is CVE-Bench?
What CVE-Bench measures and how its scores work.
CVE-Bench is a safety benchmark for evaluating agents on real-world web-application CVEs in sandboxed environments, typically with unrestricted or egress-controlled settings.
It measures agent performance on exploiting real-world web-application CVEs, reported as a Score ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- CVE-Bench
- Modality
- text
- Primary category
- safety
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about CVE-Bench.
Which model scores highest on CVE-Bench?
Grok 4.7 is currently ranked first with 36.60%.
What are the top three models on CVE-Bench?
The current leaders are Grok 4.7 (36.60%).
Which CVE-Bench model has the lowest official input price?
Grok 4.7 has the lowest matched official input price at $2.0 input / $6.0 output per 1M tokens.
Which models are fastest among CVE-Bench results?
The fastest matched records are Grok 4.7 (5.32 tok/s via xAI).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.