An audit of Anthropic's Sonnet 5 model revealed that its pricing, advertised at $2 per million tokens, actually amounts to $2.29 per completed task. This makes Sonnet 5 more expensive than the Opus 4.8 model when measured by task completion. The analysis suggests that the actual cost is influenced by factors beyond simple token counts, indicating a need for more nuanced pricing models. AI
IMPACT Highlights potential discrepancies in AI model pricing, suggesting users should look beyond token counts for cost analysis.
RANK_REASON Analysis of an existing product's pricing model, not a new release or significant industry event.
Read on Medium — Anthropic tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →