Is There a Claude Sonnet 5? Anthropic's 2026 Lineup

If you’ve searched for “Claude Sonnet 5,” you’re not alone — and you’ve probably come away confused. So let’s clear it up directly: as of mid-2026, there is no model called Claude Sonnet 5. The current Sonnet is Claude Sonnet 4.6, and the newest, most capable model Anthropic has released is Claude Fable 5 — a new flagship line, not a Sonnet.

That gap between what people search for and what actually ships is worth a quick tour, because Anthropic’s naming has more moving parts than it used to.

How Anthropic names its models

For years, the lineup followed a tidy pattern: three tiers, each with a version number.

• Haiku — the fastest, cheapest tier, for simple high-volume tasks.
• Sonnet — the balanced tier, the everyday workhorse.
• Opus — the most capable tier, for the hardest problems.

So “Sonnet 5” would simply be the next major version of the Sonnet tier. It’s a perfectly reasonable thing to expect — it just hasn’t been released. The latest Sonnet is 4.6. What has changed the picture is that Anthropic added a name above Opus entirely: Fable 5, its new top-end model for frontier reasoning and long-horizon agent work.

The actual 2026 lineup

Here’s the current family, cheapest to most capable:

Model	Tier	Context	Max output	Input $/1M	Output $/1M
Claude Haiku 4.5	Fast / cheap	200K	64K	$1.00	$5.00
Claude Sonnet 4.6	Balanced	1M	64K	$3.00	$15.00
Claude Opus 4.8	High capability	1M	128K	$5.00	$25.00
Claude Fable 5	Frontier flagship	1M	128K	$10.00	$50.00

A few things stand out. Every current model except Haiku now ships with a 1-million-token context window. All of them support adaptive thinking, where the model decides how much to reason rather than you setting a fixed budget. And the price ladder is steep at the top: Fable 5 costs twice what Opus 4.8 does, which is the whole point of having both.

So which one do you actually want?

It depends on why you came looking for “Sonnet 5” in the first place:

• You want the most capable model, full stop. That’s Claude Fable 5 — built for autonomous, multi-step agent work. Opus 4.8 is the strong, more affordable runner-up.
• You want the best balance of speed, cost, and smarts. That’s the role Sonnet has always played, and Sonnet 4.6 plays it now. It’s the right default for most applications.
• You want cheap and fast at scale. Haiku 4.5 handles classification, extraction, and routing without burning budget.

If you’ve been running an older Sonnet, the upgrade path is just Sonnet 4.6 — same tier, newer model, with adaptive thinking and the larger context window.

Will there be a Sonnet 5?

Anthropic doesn’t pre-announce its roadmap, so anyone claiming a date for “Sonnet 5” is guessing. The tier-and-version naming makes a future Sonnet 5 plausible eventually — but today, betting your project on a model that hasn’t shipped is a mistake. Build against what exists: Sonnet 4.6 for balance, Opus 4.8 and Fable 5 for capability.

The takeaway

There’s no Claude Sonnet 5 yet. The latest Sonnet is 4.6, and the new top-tier model is Fable 5. If you were chasing “the newest, smartest Claude,” Fable 5 is the answer; if you wanted “the reliable everyday model,” Sonnet 4.6 is it. For the bigger picture on where these models fit in real workflows, see our look at the state of AI coding assistants in 2026 and how open-source models are closing the gap.