Claude Sonnet 5 — the new default Sonnet, priced to undercut itself for two months

Anthropic shipped Claude Sonnet 5 on June 30 — the successor to Sonnet 4.6, GA across the Claude API (`claude-sonnet-5`), Bedrock, Vertex, and Foundry, and the new default on claude.ai Free/Pro. It keeps the 1M-token context window and is introductory-priced at $2 / $10 per 1M (input / output) through August 31 2026, reverting to the same $3 / $15 Sonnet 4.6 carried on September 1. Like every hosted frontier model, it changes no local planner pick — but it does reset the cloud-cost baseline our calculator compares against, so we updated that.

Verdict: Anthropic's new mid-tier workhorse — GA, 1M context, intro-priced below Sonnet 4.6; hosted-only, so it changes no local pick

The take

The facts, verified against anthropic.com/news/claude-sonnet-5 and platform.claude.com (both dated 2026-06-30): Sonnet 5 is the current Sonnet-tier model, 1M context, up to 128K output (300K on the Batch API beta). Pricing is $2 input / $10 output per 1M as an introductory rate through Aug 31 2026, then $3 / $15 — the exact list price Sonnet 4.6 carried, so the long-term cloud math is unchanged; the next two months are just cheaper. Standard Anthropic prompt caching (cache reads at 0.1×) and the 50% Batch discount apply. Opus 4.8 ($5/$25) and Haiku 4.5 ($1/$5) are unchanged and unaffected; this is purely a Sonnet-line refresh.

Why it matters here: Sonnet is the tier most people actually pay for when they reach for cloud — fast, capable, high-volume — so it is the most relevant cloud comparator for the "is local worth it?" question. The intro discount tilts the short-term math toward cloud for moderate usage, which is worth being honest about; the break-even on owning hardware stretches a little until Sep 1, then snaps back to where it was.

Our call: no local-pick change — Sonnet 5 is hosted-only and nothing about the local landscape moved. We replaced "Claude Sonnet 4.6" with "Claude Sonnet 5" in the cost calculator at the current $2/$10 introductory rate (with a note that it rises to $3/$15 on Sep 1), so the comparison reflects what you would actually pay this month. If you run a local rig today, Sonnet 5 is the cloud option to benchmark your own outputs against; it does not change what you should run on your own hardware.

Where this fits

Models: GLM-5.1 · Kimi K2.6 · DeepSeek V4-Pro · Qwen3-Coder-30B-A3B

Hardware: NVIDIA DGX Spark · Mac Studio M3 Ultra 96 GB · NVIDIA RTX 5090

Sources

Next step

Try this in the planner→