Vals Time Horizon Index: Kerbal Space Program (Time Horizon Index: KSP)
We show this table for reference; we do not rank on it.
A Vals AI agent benchmark that gives each system five days to build and run a space program in Kerbal Space Program.
KSP progress on Time Horizon Index: KSP — September 14, 2026
We mirror the published ksp progress view for Time Horizon Index: KSP. GPT-6 Astra leads the public snapshot at 90.500%, followed by Claude Fable 5.1 (63.333%) and GPT-5.6 Sol (23.833%). We do not use these results to rank models overall.
GPT-6 Astra
OpenAI
max reasoning
Claude Fable 5.1
Anthropic
GPT-5.6 Sol
OpenAI
max reasoning
11 modelsAgenticCurrentDisplay onlyUpdated September 14, 2026
KSP progress table (11 models)
ScoreHow Time Horizon Index: KSP is shown here
BenchLM mirrors the public Vals AI Time Horizon Index: KSP leaderboard captured from https://www.vals.ai/benchmarks/time_horizon_index and updated by Vals on September 14, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.
Time Horizon Index: KSP is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.
Snapshot
The published Time Horizon Index: KSP snapshot places GPT-6 Astra first at 90.500%. The third row is 66.667 points behind. The broader top-10 range is 84.667 points, so the table still separates the published systems.
11 models have been evaluated on Time Horizon Index: KSP. The benchmark falls in the Agentic category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Time Horizon Index: KSP is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Time Horizon Index: KSP
Year
2026
Tasks
30 progressively harder Kerbal Space Program missions
Format
Mission-ladder progress with partial credit
Difficulty
Long-horizon autonomous computer use and planning
The first Time Horizon Index uses a 30-rung mission ladder and awards partial progress within a rung. We mirror the four launch rows and progress-efficiency scores. Each result comes from one five-day run with a computer-use harness, so the small system-level table remains display only.
Freshness and provenance
Version
Time Horizon Index: KSP 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 Time Horizon Index: KSP measure?
A Vals AI agent benchmark that gives each system five days to build and run a space program in Kerbal Space Program.
Which model leads the published Time Horizon Index: KSP snapshot?
GPT-6 Astra currently leads the published Time Horizon Index: KSP snapshot with 90.500% ksp progress. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on Time Horizon Index: KSP?
The September 14, 2026 snapshot contains 11 AI models.
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 2,000+ readers.
One email each week. Unsubscribe anytime.