Anthropic's new Claude Sonnet 5 model, while competitively priced and performing well on agent-based tasks, has nearly doubled the real costs for users. This increase is due to a significant rise in token consumption per task, approximately 40% higher than its predecessor, a practice The Decoder notes is becoming a pattern for Anthropic. Despite unchanged list prices, the effective cost of using Sonnet 5 is substantially higher, impacting its overall value proposition. AI
IMPACT This pricing strategy may influence user adoption and cost management for AI models, prompting closer scrutiny of token efficiency.
RANK_REASON This is a product update for an existing model, not a frontier release, and focuses on pricing and cost implications rather than a novel capability.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
