Epoch AI - Spireclimb: leaderboard
Epoch AI's computer-use evaluation of Slay the Spire 1 and 2 at maximum difficulty. Models act through screenshots, keyboard and mouse, with the game wiki available. Score is the equally weighted mean of both games' win percentages; higher is better. Source: epoch.ai/benchmarks/spireclimb.
Metric: Score. Source: epoch.ai. Saturation forecast: Around 2028. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (Max) | 25.67 | #5 (GPT-6 Astra) |
| 2 | GPT-6.1 Sol (Max) | 24.67 | #8 (GPT-6.1 Sol) |
| 3 | GPT-6 Sol (Max) | 7.5 | #14 (GPT-6 Sol) |
| 4 | Claude Opus 5.5 (Max) | 3.33 | #3 (Claude Opus 5.5) |
| 5 | Claude Fable 5.1 (Max) | 2 | #9 (Claude Fable 5.1) |
| 6 | Claude Sonnet 5.5 (Max) | 1.33 | #10 (Claude Sonnet 5.5) |
| 7 | GPT-6 Luna (Max) | 0 | #52 (GPT-6 Luna) |
No result here: #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=epoch-ai-spireclimb · How It Works · Data refreshed daily, snapshot 2026-10-16.