[AINews] Sonnet 5 today, and Fable 5 tomorrow

At standard pricing, Sonnet 5 costs $2.29 per Intelligence Index task—about 2x what Sonnet 4.6 costs and roughly 15% more than Opus 4.8—because it uses more tokens per task despite a lower per-token price.

Artificial Analysis calculated that Sonnet 5's higher token usage per task makes it cost more in practice than even Opus 4.8—$2.29 per Intelligence Index task, roughly 2x Sonnet 4.6's cost—undercutting the appeal of its lower sticker price.

explains mechanism · 1

Sonnet 5 could launch as the widely accessible but weaker companion model, while a stronger Fable 5 stays gated behind more restrictive, regionally limited access, especially in Europe.

Ahead of launch, commentator kimmonismus worried that Sonnet 5 would be the open, lower-tier release while a more powerful Fable 5 stayed locked behind identity verification and regional restrictions, particularly affecting Europe.

gives example · 1

Sonnet 5 offers top-tier coding and tool-use performance at Sonnet-level pricing, includes a 1-million-token context window, and is now the default model in Claude Code for Pro users across the Claude Platform, API, and Managed Agents.

Anthropic's developer account positioned Sonnet 5 as delivering frontier-quality coding and agentic tool-use at Sonnet's lower price point, with a 1M-token context window, becoming the new default model in Claude Code and across the Claude Platform.

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.