Each Time AI Gets Smarter, We Change the Definition of Intelligence
Definitions and benchmarks of humanlike intelligence shift as AI improves, complicating contractual rights and independent verification of any AGI claim.
OpenAI says GPT-5 stacks up to humans in a wide range of jobs | TechCrunch
GDPval benchmark evaluates AI models across 44 occupations in nine GDP-contributing industries; GPT-5-high matched or exceeded experts 40.6% of the time.
Kaggle Introduces Game Arena to Benchmark AI Models in Strategic Games
Kaggle and Google DeepMind launched Kaggle Game Arena to benchmark AI decision-making by running all-play-all strategy game competitions with open-source environments.
LM Arena, the organization behind popular AI leaderboards, lands $100M | TechCrunch
LM Arena has become an essential crowdsourced benchmarking project for AI labs, raising $100 million in seed funding to further its mission of evaluating AI models.
OpenAI's o3: AI Benchmark Discrepancy Reveals Gaps in Performance Claims
The performance of OpenAI's o3 model on benchmarks significantly differed from earlier claims, revealing the complexity and variability in AI evaluations.
AI benchmarking platform Chatbot Arena forms a new company | TechCrunch
Chatbot Arena is forming a company called Arena Intelligence Inc. to enhance its benchmarking capabilities significantly while maintaining neutrality in AI testing.
Last week, a post on X claimed Google's Gemini model surpassed Anthropic's Claude model in Pokemon, stirring controversy over AI benchmarks and implementation.
