Kimi K3, and what we can still learn from the pelican benchmark

This article discusses the release of Moonshot AI's Kimi K3, a 2.8 trillion parameter model, comparing its performance against leading models like Claude Opus 4.8, GPT-5.5, and Claude Fable 5. It covers benchmark results from Artificial Analysis, pricing details ($3/$15 per million tokens), and the model's leading position on Arena.ai's Frontend Code arena. The author also reflects on the 'pelican riding a bicycle' SVG test as an informal benchmark, noting its diminishing correlation with model quality over time. The article is relevant to IT/technology as it covers AI model advancements, performance metrics, and cost analysis.