Sonnet 5 - its updated tokenizer maps the same text to more tokens (roughly 1.0–1.35× depending on content), so cost per task can be higher.

So they didn’t raise the price but it costs more? Seems fishy to me, like their “new tokenizer” is trying to get more cash. Also love the “first hit is free reduced” “Claude Sonnet 5 is available across your organization today. It's our most agentic Sonnet model yet, able to plan, use tools, and run autonomously on complex work — with performance close to Claude Opus 4.8 at a lower price and substantial gains over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work. What your members start on depends on their seat type: Users on Standard seats now default to Sonnet 5 across chat, Cowork, and Claude Code for new conversations. Users on Premium seats continue to default to Opus 4.8 across chat, Cowork, and Claude Code, with Sonnet 5 available in the model picker. It lists at the same price as Sonnet 4.6 ($3/$15 per million input/output tokens), but its updated tokenizer maps the same text to more tokens (roughly 1.0–1.35× depending on content), so cost per task can be higher. To offset the tokenizer's effect on cost per task during rollout, introductory pricing of $2/$10 per million input/output tokens applies through August 31, 2026. Standard pricing resumes September 1st”