Near-Opus quality at the same permanent $2/$10 rate Sonnet 5 held. The default Claude for most production work and Claude's free-tier model.
Claude Sonnet 5.5 replaced Sonnet 5 on September 28, 2026, across the Claude API, claude.ai, AWS, Google Cloud, and Microsoft Foundry. Pricing is unchanged at $2/$10 per 1M tokens (cache read $0.20, cache write $2.50). Sonnet 5 is now a legacy model — still available, with retirement not sooner than September 2027 — while Sonnet 5.5 takes over as the current, recommended Sonnet.
Sonnet 5.5 carries over Sonnet 5's price exactly: $2 / 1M input and $10 / 1M output, with cache reads at $0.20/1M and cache writes at $2.50/1M. Anthropic made this a routine successor — no action needed from existing integrations.
Claude Sonnet 5.5, released September 28, 2026, replaces Sonnet 5 (June 30, 2026) at the identical $2 input / $10 output per-1M-token price — the same permanent rate Anthropic locked in for Sonnet 5 on August 10, 2026. Cache pricing also carries over unchanged: $0.20/1M cache reads, $2.50/1M cache writes. It's now available across the Claude API, claude.ai, Claude Code, AWS, Google Cloud, and Microsoft Foundry.
The Sonnet pitch is otherwise unchanged in shape: most of Opus's capability at a meaningful discount, with better latency. Anthropic's own models-overview documentation now lists Sonnet 5.5 — alongside Fable 5.1, Opus 5.5, and Haiku 4.5 — as the current generation, with Sonnet 5 moved to a "legacy models (still available)" list alongside Opus 5 and older Opus releases. Legacy doesn't mean retired: Anthropic commits to keeping Sonnet 5 available no sooner than September 2027.
Sonnet 5.5 is a strong all-rounder: writing, code, vision, and structured extraction. It's the right model for production chat agents and coding assistants because answers stay composed under load, and tool use is reliable across multi-turn loops. As with the Sonnet 5 → 5.5 transition, the model card emphasizes gains in agentic execution and adaptive thinking (the effort parameter defaults to High on the API, Medium in Claude Code and the apps).
Where it still trails the Opus tier (Opus 5 and Opus 5.5): the hardest novel reasoning, agentic coding, and the most demanding long-document synthesis. Where it trails GPT-5.5: outright token throughput.
| Model | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| Claude Sonnet 5.5 | $2 | $10 | 1M |
| Claude Opus 5.5 | $4 | $20 | 1M |
| Claude Opus 5 | $5 | $25 | 1M |
| Claude Haiku 4.5 | $1 | $5 | 200K |
| GPT-5.4 | $2.50 | $15 | 272K |
Sonnet 5.5's $2/$10 rate makes it a clear mid-tier value pick — cheaper on output than GPT-5.4, with a full 1M context window. Opus 5.5 and Opus 5 cost 2–2.5x more for the highest-stakes work; Haiku 4.5 is half the price for narrower tasks.