AI

Topic

#benchmarks

2 articles

Gemini Ultra 2 Tops Every Major Benchmark — What It Means
Google AI / DeepMindFeb 24, 2026

Gemini Ultra 2 Tops Every Major Benchmark — What It Means

Google's latest flagship model outperforms competitors on MMLU, HumanEval, and a new multimodal reasoning suite. Here's the full breakdown.

2 min read
OpenAI's o3 Is Here: A Reasoning Model That Thinks Before It Speaks
OpenAIFeb 22, 2026

OpenAI's o3 Is Here: A Reasoning Model That Thinks Before It Speaks

OpenAI's o3 model introduces variable compute at inference time — letting it 'think longer' on hard problems. Early results suggest a step-change in complex reasoning.

2 min read