Inside DiG-bench: 70 Handcrafted Text Games to Test AI Exploration
jcrwhittington · x · 2026-08-12
DiG-bench consists of 70 self-contained, distinct text-based game worlds where players (or AIs) must discover hidden rules and win conditions through experimentation.
The games are handcrafted, novel, and not available on the internet (21 public, 49 private for secure evaluation). With 7 difficulty tiers, all games are hard but beatable by humans on their first attempt, serving as a strictly calibrated benchmark for frontier AI exploration capabilities.
Related event: DiG-bench: 70 Text Games Expose Frontier AI's Discovery Gaps(14 posts)→
More from Research
- DIY LEGO bench tests AI brick-building, GPT-6 Astra beats Opus — TheLegoless · 2026-10-16
- DIRECT: retrieval-based next-question suggestions for conversational agents — burkov · 2026-10-16
- AI-Driven Timing-Aware Power Optimization Uses Learned Ranking With Baseline-Preserving Repair — Amit Sharma · 2026-10-16
- Waypoint-OPD: Alibaba Gates On-Policy Distillation via Early-Late Teacher Shift Agreement — Bingxi Hou · 2026-10-16
- MJev-Spatial: Single-Pass Classification Beats Autoregressive VLMs on Spatial Reasoning — mindlab-research · 2026-10-16
- MLExpert: 35B MoE Agent Trained to Autonomously Handle Machine Learning Engineering — Shangheng Du · 2026-10-16