Knowledge · Benchmark profile
A cybersecurity benchmark for evaluating LLM agents on full-control V8 exploit synthesis using 16 measured exploit capability flags.
Data verified 27 Jul 2026 · Methodology 1.8.0
Benchmark score on ExploitBench v8-bench
GPT-5.6 Sol leads at 73.5%, followed by GPT-5.6 Terra (52.9%) and GPT-5.6 Luna (33.2%).
4 ranked models · higher is better · labels show rank and score
| Model | Rank | Provider | Score |
|---|---|---|---|
| GPT-5.6 Sol | #1 | OpenAI | 73.5% |
| GPT-5.6 Terra | #2 | OpenAI | 52.9% |
| GPT-5.6 Luna | #3 | OpenAI | 33.2% |
| Kimi K3 | #4 | Moonshot AI | 32% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | GPT-5.6 Sol gpt-5.6-sol | OpenAI | closed | Reference only | 73.5% |
| #2 | GPT-5.6 Terra gpt-5.6-terra | OpenAI | closed | Estimated reference | 52.9% |
| #3 | GPT-5.6 Luna gpt-5.6-luna | OpenAI | closed | Estimated reference | 33.2% |
| #4 | Kimi K3 kimi-k3 | Moonshot AI | closed | Reference only | 32% |
The top of this snapshot is led by GPT-5.6 Sol at 73.5%; third place is 40.3 points behind. The top-4 spread is 41.5 points.
About ExploitBench v8-bench
A cybersecurity benchmark for evaluating LLM agents on full-control V8 exploit synthesis using 16 measured exploit capability flags. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A cybersecurity benchmark for evaluating LLM agents on full-control V8 exploit synthesis using 16 measured exploit capability flags.
GPT-5.6 Sol by OpenAI currently leads with 73.5%.
4 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
2026 · 21 results · Reference
Artificial Analysis MMLU-Pro2026 · 8 results · Reference
Massive Multitask Language Understanding2020 · 16 results · Reference
Graduate-Level Google-Proof Q&A2023 · 140 results · Reference
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines2025 · 38 results · Reference
Massive Multitask Language Understanding Professional2024 · 85 results · Reference
