The Last Ones Cyber Range Completion Rate (The Last Ones completion)
We show this table for reference; we do not rank on it.
Share of runs that completed the 32-step cyber range within the 100-million-token limit.
Benchmark score on The Last Ones completion — November 11, 2026
We compile the The Last Ones completion rows from benchmark-owner or independent runs and provider self-reports. GPT-5.6 Sol leads the table at 70%, followed by Claude Mythos 5 (60%) and Kimi K3 (10%). We do not use these results to rank models overall.
GPT-5.6 Sol
OpenAI
Claude Mythos 5
Anthropic
Kimi K3
Moonshot AI
3 modelsAgenticCurrentDisplay onlyUpdated November 11, 2026
Benchmark score table (3 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | GPT-5.6 SolOpenAI | 70% | Not reported | Closed |
| 2 | Claude Mythos 5Anthropic | 60% | Not reported | Closed |
| 3 | Kimi K3Moonshot AI | 10% | Not reported | Open |
Among the reported The Last Ones completion rows, GPT-5.6 Sol is first at 70%. The third row is 60 points behind. The broader top-10 range is 60 points, so the table still separates the published systems.
3 models have been evaluated on The Last Ones completion. The benchmark falls in the Agentic category. We keep these results separate from the weighted global scoring system, so they remain source-specific evidence. The Last Ones completion is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About The Last Ones completion
Year
2026
Tasks
10 long-horizon runs
Format
Completion rate
Difficulty
Long-horizon cyber operations
The preliminary assessment reports completed runs out of ten under a 100-million-token cap. BenchLM expresses that ratio as a percentage and keeps it display only.
Freshness and provenance
Version
The Last Ones completion 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 The Last Ones completion measure?
Share of runs that completed the 32-step cyber range within the 100-million-token limit.
Which model scores highest on The Last Ones completion?
GPT-5.6 Sol by OpenAI currently leads with a score of 70% on The Last Ones completion.
How many models are evaluated on The Last Ones completion?
3 AI models have published results on The Last Ones completion in the BenchLM catalog.
Compare top models on The Last Ones completion
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.