APIUp to 25% cheaper than official pricesTry the API →
Hermes
APIModelsCases
Back to News
newsBenchmarksReasoning

Tencent's ExplorationBench Forces AI to Discover Hidden Rules Through Active Experimentation

Tencent, Fudan, and Tsinghua built two alien worlds with rewritten rules to test whether frontier models can discover new knowledge by probing an environment.

Tencent's ExplorationBench Forces AI to Discover Hidden Rules Through Active Experimentation
Source
Tencent Hy
Published
Author
AlphaSignal Newsroom
Read
1 min read

Tencent, Fudan, and Tsinghua built two alien worlds with rewritten rules to test whether frontier models can discover new knowledge by probing an environment.

Reporting is indexed from AlphaSignal. Rights remain with the original publisher and cited sources.

Read original report

Next reads

Epoch's SpireClimb Shows GPT-6 Astra Winning 98% of Slay the Spire Runs
Epoch AI · news

Epoch's SpireClimb Shows GPT-6 Astra Winning 98% of Slay the Spire Runs

Stanford's Priced Guidance Framework Measures How Hard AI Finds Research Ideas
AlphaSignal · paper

Stanford's Priced Guidance Framework Measures How Hard AI Finds Research Ideas

OpenAI's GPT-6.1 Sol Cuts Benchmark Costs 76% but Runs 3x Slower
Vals AI · news

OpenAI's GPT-6.1 Sol Cuts Benchmark Costs 76% but Runs 3x Slower