Atomic Vulnerability Research and Exploitation (Atomic vulnerability research)
We show this table for reference; we do not rank on it.
Irregular's domain-level evaluation of vulnerability research and exploitation capability.
Benchmark score on Atomic vulnerability research — November 13, 2026
We compile the Atomic vulnerability research rows from benchmark-owner or independent runs. GPT-6 Astra leads the table at 100%, followed by GPT-5.5 (92%) and GPT-5.6 Sol (91%). We do not use these results to rank models overall.
GPT-6 Astra
OpenAI
Run by Atomic vulnerability researchGPT-5.5
OpenAI
Run by Atomic vulnerability researchGPT-5.6 Sol
OpenAI
Run by Atomic vulnerability research3 modelsAgenticCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (3 models)
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | GPT-6 AstraOpenAIRun by Atomic vulnerability research | 100% | Run by Atomic vulnerability research | Not reported | Closed |
| 2 | GPT-5.5OpenAIRun by Atomic vulnerability research | 92% | Run by Atomic vulnerability research | Not reported | Closed |
| 3 | GPT-5.6 SolOpenAIRun by Atomic vulnerability research | 91% | Run by Atomic vulnerability research | Not reported | Closed |
- Lab-reported
- Benchmark-run
- Independent
Among the reported Atomic vulnerability research rows, GPT-6 Astra is first at 100%. The third row is 9 points behind. The broader top-10 range is 9 points, so many of the published results sit in a relatively narrow band.
3 models have been evaluated on Atomic vulnerability research. The benchmark falls in the Agentic category. We keep these results separate from the weighted global scoring system, so they remain source-specific evidence. Atomic vulnerability research is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Atomic vulnerability research
Year
2026
Tasks
Atomic cyber tasks
Format
Domain average
Difficulty
Vulnerability research and exploitation
This Atomic domain average isolates vulnerability research and exploitation. BenchLM stores it as an independent, display-only signal.
Atomic vulnerability research is one of the cyber benchmarks we track. The cybersecurity benchmark comparison sets it beside the others, one table per security job.
Freshness and provenance
Version
Atomic vulnerability research 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Atomic vulnerability research measure?
Irregular's domain-level evaluation of vulnerability research and exploitation capability.
Which model scores highest on Atomic vulnerability research?
GPT-6 Astra by OpenAI currently leads with a score of 100% on Atomic vulnerability research.
How many models are evaluated on Atomic vulnerability research?
3 AI models have published results on Atomic vulnerability research in the BenchLM catalog.
Compare top models on Atomic vulnerability research
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.