KradleBench launches as a dynamic multi-agent benchmark pitting LLMs, world models and humans against each other
lmoroney · x · 2026-11-19
Kradle AI, founded by ex-Googlers James Tamplin and Kelmoujahid, announced KradleBench — a dynamic, multi-agent benchmark designed for both AIs and humans:
- Participants include LLMs, world models, and humans, compared in real-world-like environments for capabilities and safety
- Aimed at generating useful signals for improving current systems
- Soon opening up for more human players and new eval proposals, teasing whether a human can dethrone "GPT 6.2 Astra"
Developers can also submit their own models for testing.
More from Research
- Shanghai AI Lab's GameRealm: 120M frames of gameplay data for visual foundation models — Shanghai-AI-Laboratory · 2026-11-19
- KradleBench Launches: Minecraft Legend SethBling Dethroned by Model Astra 6.2 — fchollet · 2026-11-19
- Transformer Lab unveils Primus, a 'society of researchers' of AI agents, and is hiring — carsonfarmer · 2026-11-19
- New LessWrong Analysis: LLMs Seek the Grader's Score, Not the Actual Goal — 1a3orn · 2026-11-19
- tabgen-eval: an open-source eval framework and benchmark for synthetic tabular data — LucaAmb · 2026-11-19
- New meta-analyses slash GMA's predictive power for job performance to ~0.2 correlation — rasbt · 2026-11-19