I Did the Math on Claude Sonnet 5. The 60% Opus Discount Is Real, But Temporary.

Anthropic made Sonnet 5 broadly available at $2/$10 per million tokens until August 31. I ran the routing math; it should be your default, not your ceiling.

---

title: "I Did the Math on Claude Sonnet 5. The 60% Opus Discount Is Real, But Temporary."

published: true

description: "Anthropic made Sonnet 5 broadly available at $2/$10 per million tokens until August 31. I ran the routing math; it should be your default, not your ceiling."

tags: ai, anthropic, claude, programming

cover_image: https://cdn.tokenmix.ai/tongyong.png

canonical_url: https://tokenmix.ai/blog/claude-sonnet-5-review-pricing-benchmark

---

Anthropic shipped Claude Sonnet 5, and the takes I saw were predictable:

"It replaces Opus."

"It is just another Sonnet refresh."

"The benchmark chart means you can route everything to it now."

Two of those are wrong. One is directionally right, but only if you care about cost per task instead of model prestige.

I spent time going through Anthropic's launch post, the Claude Platform docs, GitHub's Copilot rollout note, and the pricing math. The conclusion I landed on is simple: **Sonnet 5 should be the default Claude model for most coding agents, but it should not be your highest-stakes escalation model.**

TL;DR

**No, Sonnet 5 does not universally replace Opus 4.8.** Anthropic says it can match Opus on some higher-effort tasks, not all tasks.
**Yes, the discount is real.** Intro pricing is $2 input / $10 output per million tokens through August 31. Opus 4.8 is $5/$25.
**The real number is 60%.** During the intro period, Sonnet 5 costs 40% of Opus 4.8, meaning a 60% discount on both input and output.
**After August 31, the math changes but still works.** Sonnet 5 moves to $3/$15, still 40% cheaper than Opus 4.8.
**My routing rule:** use Sonnet 5 for the first pass, Opus 4.8 for escalation, and Fable 5 only when the task justifies frontier-tier cost.

What actually shipped

Anthropic launched Claude Sonnet 5 on June 30, 2026.

The important part is not just the model. It is the availability.

Sonnet 5 is available across Claude Free, Pro, Max, Team, Enterprise, Claude Code, Claude Cowork, and the Claude Platform API, according to [Anthropic's launch post](https://www.anthropic.com/news/claude-sonnet-5). GitHub also made Sonnet 5 generally available in Copilot on June 30, which means this model landed directly inside developer workflows, not just API dashboards.

That matters because the frontier tier is noisy right now:

| Model / product | Current reality |

| --------------- | ----------------------------------------------- |

| Claude Fable 5 | Back online, but expensive and policy-sensitive |

| Claude Mythos 5 | Narrower access |

| GPT-5.6 | Gated preview, not broadly available |

| Gemini 3.5 Pro | Reported July target, not public API yet |

| Claude Sonnet 5 | Broadly available now |

This is why I care about Sonnet 5 more than the louder frontier-model drama.

It is the model developers can actually use this week.