Vals AI Time Horizon Index: leaderboard
Vals AI Kerbal Space Program agent benchmark: how far a model progresses through a 30-mission space program (rocket design, orbital mechanics) in a 120-hour window; each mission rung is 3.33% progress.
Metric: Progress (%). Source: www.vals.ai. Saturation forecast: Around February 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | 91.33 | #4 (Claude Opus 5.5) |
| 2 | GPT-6 Astra (Max) | 90.5 | #6 (GPT-6 Astra) |
| 3 | Claude Fable 5.1 (Max) | 63.33 | #9 (Claude Fable 5.1) |
| 4 | Gemini 4 Argon (High) | 52.5 | #8 (Gemini 4 Argon) |
| 5 | GPT-5.6 Sol (Max) | 23.83 | #20 (GPT-5.6 Sol) |
| 6 | Claude Opus 5 (Max) | 18.83 | #13 (Claude Opus 5) |
| 7 | Gemini 3.8 Flash (High) | 18.83 | #24 (Gemini 3.8 Flash) |
| 8 | Claude Opus 4.8 (Max) | 11.83 | #36 (Claude Opus 4.8) |
| 9 | Kimi K3 (Max) | 10.5 | #32 (Kimi K3) |
| 10 | GPT-5.5 (xHigh) | 8.33 | #30 (GPT-5.5) |
| 11 | Grok 4.5 (High) | 6.33 | #48 (Grok 4.5) |
| 12 | Grok 4.6 (High) | 5.83 | #27 (Grok 4.6) |
| 13 | Muse Spark 1.1 (xHigh) | 1.67 | #55 (Muse Spark 1.1) |
No result here: #7 GPT-6.1 Sol, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-time-horizon-index · How It Works · Data refreshed daily, snapshot 2026-10-24.