MechanismArticle

The new tokenizer makes Sonnet 5 about 1.4x more expensive for English text and 1.33x more expensive for Spanish, while Simplified Mandarin costs about the same.

Simon Willison flagged that Sonnet 5's new tokenizer quietly raises effective costs for some languages—about 1.4x for English and 1.33x for Spanish—while leaving Simplified Mandarin roughly unchanged, a detail obscured by comparing list prices alone. ✦ AI generated

Simon Willison · Latent Space · 2026-07-01 · original ↗

Export .md Export .txt

One underappreciated technical detail was the tokenizer/effective billing behavior. Simon Willison noted the new tokenizer makes Sonnet 5 ~1.4x more expensive for English, ~1.33x for Spanish, and roughly the same for Simplified Mandarin

Read full article ↗excerpt · fair-use quotation

✳ Sonnet 5 ◉ Simon Willison

Related moments

Mechanism◆Audio · 0:00 · 2m

If you don't batch together many concurrent users on a GPU cluster, the cost per token can be a thousand times worse than if you do batch them, because batching amortizes the fixed cost of fetching model weights from memory across many tokens at once.

Reiner Pope · Dwarkesh Podcast

DataArticle

The full multi-agent sqlite-utils 4.0rc2 development session cost an estimated $149.25 in unsubsidized API costs, which is why Simon Willison upgraded to the $200/month Claude Max plan ahead of the July 7th API-pricing cutover.

Simon Willison · Simon Willison's Weblog

PredictionArticle

GLM-5.2's diffusion while Claude Fable remains export-banned deals a severe economic blow to frontier labs, letting open models erode the lower-margin business those labs need while they push into higher-margin frontier domains.

Nathan Lambert · Interconnects