GPT-5 launched yesterday. 94.6% on AIME 2025. 74.9% on SWE-bench.
As we approach the upper bounds of these benchmarks, they die.
What makes GPT-5 and the next generation of models revolutionary isn’t their knowledge. It’s knowing how to act. For GPT-5 this happens at two levels. First, deciding which model to use. But second, and more importantly, through tool calling.
We’ve been living in an era where LLMs mastered knowledge retrieval & reassembly.
Recent Stories
Startup Funding: Q4 2025
More and bigger funding rounds for AI chips and AI for making chips; 75 companies raise $3 billion.
Jan 19, 2026Policymakers And Lawmakers Eyeing The Use Of AI As A Requisite First-Line For Mental Health Gatekeeping And Therapy Intervention
Should we use AI to pre-screen whether people can see a human therapist? And should we use AI to do initial intervention? It's a huge controversy. An AI Insider scoop.
Jan 19, 2026FedEx CEO explains why regular humanoid robots can’t get the job done
Raj Subramaniam said "super humanoid robots" with a "couple of elbows" would be better at automating deliveries.