Anthropic’s new Claude Sonnet 5 model has shown impressive performance, beating the pricier Opus 4.8 on some tasks, but its increased token consumption makes it more expensive to use, with costs nearly doubling despite unchanged token prices. The model’s improved performance comes at a cost, with a significant jump in token usage, making it less competitive than it initially seems. This trend of hiding price increases behind unchanged token rates is a concern for developers and users, and it’s not the first time Anthropic has done this, with previous models also showing increased costs without explicit price hikes. Why it matters: this subtle price increase strategy could impact the adoption and accessibility of AI models, especially for smaller developers and businesses.