K2-Thinking 开源，支持 300 步工具调用（附：绝世好 prompt）
"AI能力评估" 相关文章
The "Length" of "Horizons"
[GDPval] Models Could Automate the U.S. Economy by 2027
What Does Your Benchmark Really Measure? A Framework for Robust Inference of AI Capabilities
GPT-5冷酷操盘，狼人杀一战封神！七大LLM狂飙演技，人类玩家看完沉默
Evaluating Prediction in Acausal Mixed-Motive Settings
Do model evaluations fall prey to the Good(er) Regulator Theorem?
Anthropic Is Going All In On Ability Without Intelligence?
The Mirror Test: How We've Overcomplicated AI Self-Recognition
The Mirror Test: How We've Overcomplicated AI Self-Recognition
The Elicitation Game: Evaluating Capability Elicitation Techniques
DrafterBench: Benchmarking Large Language Models for Tasks Automation in Civil Engineering
Interpreting the METR Time Horizons Post
Recent AI model progress feels mostly like bullshit
The Elicitation Game: Evaluating capability elicitation techniques
These researchers used NPR Sunday Puzzle questions to benchmark AI ‘reasoning’ models
Understanding Benchmarks and motivating Evaluations
“人类终极考试”基准测试发布：顶级 AI 系统表现惨淡，回答准确率均未超 10%
