Slay the Spire as a long-horizon reasoning benchmark: 11 frontier models tested
JenniferHli · x · 2026-10-20
Preference Model proposes using the roguelike deck-builder Slay the Spire as a long-horizon reasoning benchmark. A good test should have clear rules, be easy to measure, and demand many deep decisions whose consequences span short and long horizons — one run consists of hundreds of consecutive decisions, with randomized events forcing on-the-fly adaptation.
Setup: 11 frontier models, each at 3 reasoning effort levels, play 5 game scenarios with 10 runs per combination. Two questions: is Slay the Spire actually a good measure of long-horizon reasoning, and which models are strongest? All runs are watchable online.
More from Models
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- TLA+ expert corrects Grok: TLC only exhausts finite state spaces, not all behaviors — tianyin_xu · 2026-10-23
- Open LLM cost calculator shows Arabic tokenization and 6x price gaps across models — Moham-saed · 2026-10-23
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23